
Task-level PR creation with formatted title, labels, structured body, and target branch.
Canonical catalog of file-collision heuristics for parallel task, story, and epic execution. Defines the File Footprint block format, the three Conflict Categories (hard / regen / soft), a hotspot list of high-contention paths, and the degrade-with-warning policy. Consumed by x-task-plan, x-story-plan, x-epic-map, x-parallel-eval, and x-dev-*-implement.
Thin orchestrator (~320 lines — story-0049-0019 refactor) that drives a story end-to-end via 4 delegated phases: Phase 0 (args via x-internal-args-normalize + context via x-internal-story-load-context + resume via x-internal-story-resume), Phase 1 (parallel planning via x-internal-story-build-plan), Phase 2 (task execution loop via x-task-implement per task, then final story PR via x-pr-create), Phase 3 (verify via x-internal-story-verify + report via x-internal-story-report + optional worktree cleanup via x-git-worktree). New EPIC-0049 flags --target-branch / --auto-merge / --epic-id propagate OO-style to x-task-implement and x-pr-create. Backward compatible: absent flags preserve legacy EPIC-0048 behavior (target=develop, auto-merge=none).
Multi-agent story planning: launches 5 specialized agents (Architect, QA, Security, Tech Lead, Product Owner) in parallel to produce a consolidated task breakdown, individual task plans, planning report, and DoR validation. Schema-aware: v1 (legacy) runs the original 6-phase flow; v2 (task-first, EPIC-0038) adds Phases 4a-4c that emit task-TASK-NNN.md + plan-task-TASK-NNN.md per task and a task-implementation-map-STORY-*.md, wiring every task through x-task-plan in parallel.
Observability principles: distributed tracing (span trees, mandatory attributes), metrics naming conventions, structured logging with mandatory fields, health checks (liveness/readiness/startup), correlation IDs, and OpenTelemetry integration.
Performance engineering patterns: profiling, benchmarking, optimization, regression detection, and memory management for {{LANGUAGE}} {{FRAMEWORK}} projects using {{BUILD_TOOL}}.
Transactional Outbox Pattern: reliable event publishing with polling publisher and CDC strategies, outbox table design, and anti-patterns.
Quarkus patterns: CDI, @ConfigMapping, Panache Repository, RESTEasy Reactive, native build constraints, and @RegisterForReflection.
Story decomposition and planning: layer-by-layer decomposition (foundation, core domain, extensions, compositions, cross-cutting), story self-containment (data contracts, acceptance criteria), dependency DAG, sizing rules, and phase computation.
Generates or updates CI/CD pipelines per project stack with actionlint validation.
Cross-cutting data modeling patterns: schema design (soft delete, temporal tables, audit trails, multi-tenant, SCD), concurrency, and test data management.
Creates a bare git branch (no worktree) from a configurable base with naming validation and idempotency. Single source of truth for branch creation logic consumed by orchestrators (x-internal-epic-branch-ensure, x-story-implement) and users.
Detects cross-epic P95 regressions and ranks top-10 slowest skills from global index.
Creates a bare git branch from a configurable base with naming validation and idempotency.
Runs integration tests covering the full request flow with a real database.
Master PR review orchestrator: specialist reviews + developer reviews + Tech Lead GO/NO-GO with auto-remediation loop.
Drives a story end-to-end via 4 phases: plan, task loop, verify, report.
Reads PR review comments and fixes actionable ones with Conventional Commits.
Implements a task via TDD Red-Green-Refactor with one atomic commit via x-commit-changes.
Orchestrates parallel story planning: 5 sibling agents (impl, tests, tasks, sec, comp).
Materializes the feature container artifact (feature-<ID>.md) from an existing Feature/spec source. Internal skill invoked by x-refine-feature Phase E.
Parses YAML frontmatter from a patterns or knowledge file and returns structured JSON with capabilities, phases, arch_questions, scoring_config, and triggers_context_packs.
Renders _TEMPLATE-*.md from a JSON payload and writes the result atomically.
Merges a single PR via gh CLI with configurable strategy and idempotent behavior.
Lightweight PO-centric business refinement of a feature or a story (value, problem, persona, scope, business AC, KPI). Writes a business verdict; never unblocks technical gates.
API specialist review: dynamically loads integration-style patterns for REST, webhook, and socket APIs.
Database specialist review: dynamically loads versioned patterns for detected DB engine and version.
Reviews event schemas, producer/consumer patterns, error handling, and DLQ readiness.
Reviews code changes for compliance with selected security frameworks and patterns.
Performance review: N+1, pools, async, pagination, cache, timeouts, circuit breakers.
Scans code and git history for leaked credentials, API keys, and tokens; SARIF output.
Audits the policy set for structural consistency and enforcement wiring; reports orphans, schema drift, and dangling skill references.
Documentation freshness gate: validates 6 dimensions (readme, api, adr, etc.) per PR.
Batch-commits planning artifacts under plans/** without the code pre-commit chain.
Unified feature entrypoint orchestrating ideate → refine → plan-arch with checkpoint/resume and persisted open questions; or legacy spec decomposition.
Detects file-collision risks between parallel work units; emits collision matrix.
Drives a story end-to-end via 4 phases: plan, task loop, verify, report.
Unified feature entrypoint orchestrating ideate → refine → plan-arch with checkpoint/resume and persisted open questions; or legacy spec decomposition.
Decomposes an implementation plan into TDD tasks (Red/Green/Refactor) from test scenarios.
Ensures feature/XXXX branch exists idempotently, locally and on origin.
Java-specific developer review: dynamically loads versioned patterns based on detected Java + Spring Boot version.
Selects canonical patterns/knowledge by phase and capability via catalog.md discovery. Aggregates arch_questions (from catalogs) and scoring_summary (from sub-patterns) for dynamic skill behaviour.
API specialist review: dynamically loads integration-style patterns for REST, webhook, and socket APIs.
Orchestrates language-specific developer reviews in parallel based on detected technologies.
Master PR review orchestrator: specialist reviews + developer reviews + Tech Lead GO/NO-GO with auto-remediation loop.
Compliance frameworks (conditionally included): GDPR, HIPAA, LGPD, PCI-DSS, SOX. Data classification, rights enforcement, processing records, international transfers, security measures, audit logging, and framework-specific requirements.
Hexagonal architecture reference: canonical package structure, dependency rules with violation examples, compilable Port/Adapter patterns, and ArchUnit boundary validation suite. Read before implementing hexagonal-style projects.
Automated profiling: detect language/runtime, select appropriate profiler, execute session, generate flamegraph, identify hotspots, and suggest optimizations referencing the performance-engineering knowledge pack.
Create Jira Stories from existing local story markdown files. Read all story files in an epic directory, map fields to Jira, create issues with parent epic link, create dependency links between stories, and sync Jira keys back to local files.
Integrates with SonarQube/SonarCloud for security hotspot tracking, quality gate enforcement, and SARIF output from findings.
Compliance frameworks (conditionally included): GDPR, HIPAA, LGPD, PCI-DSS, SOX. Data classification, rights enforcement, processing records, international transfers, security measures, audit logging, and framework-specific requirements.
Architecture pattern references: microservice, resilience, data, integration, and architectural patterns (saga, outbox, circuit breaker, CQRS, event sourcing, and more).
Hexagonal architecture reference: canonical package structure, dependency rules with violation examples, compilable Port/Adapter patterns, and ArchUnit boundary validation suite. Read before implementing hexagonal-style projects.
CQRS/Event Sourcing patterns: write/read model separation, command bus, event store, aggregate with event sourcing, projections, snapshot policy, and dead letter handling.
Diagnoses errors, stacktraces, build failures, and unexpected behavior. Systematic approach: reproduce, locate, understand, fix, verify. Use whenever something fails: compilation errors, test failures, runtime exceptions, coverage gaps, or performance issues.
Audits compliance of all project rules AND knowledge packs against source code. Launches parallel subagents (one per rule/knowledge-pack) for scanning, then aggregates into a unified report with severity classification and story suggestions.
Scans code and git history for leaked credentials, API keys, tokens, and secrets. Produces SARIF output with scoring and baseline support.
Scans Docker images for CVEs and Dockerfile best practices violations. Uses Trivy, Grype, or Snyk Container. Produces SARIF output with scoring.
DevOps specialist review: validates Dockerfile, container security, CI/CD pipeline, resource limits, health probes, graceful shutdown, and deployment configuration.
Validates REST API endpoints for RFC 7807 error responses, pagination, URL versioning, OpenAPI documentation, status codes, and DTO patterns.
Sets up the local development environment including container orchestrator, database, and build tools.
PCI-DSS compliance review with 25-point checklist for code changes involving payment card data. Produces per-point PASS/FAIL report with remediation.
Incrementally updates the service architecture document with changes from architecture plans. Adds new components, integrations, flows, and ADR references without rewriting existing content. Use after implementation to keep architecture documentation current.
Batch-commits planning artifacts under plans/** without the code pre-commit chain.
Audits dependencies for CVEs, outdated versions, and license issues per stack.
Adds a newly discovered story to an already-refined feature; validates and proposes feature/implementation-map updates.
Audits the policy set for structural consistency and enforcement wiring; reports orphans, schema drift, and dangling skill references.
Generates plan-task-*.md with TDD cycles in TPP order, file-impact, and exit criteria.
Runs integration tests covering the full request flow with a real database.
Creates a bare git branch from a configurable base with naming validation and idempotency.
Documentation automation v2: stack-aware generation from documentation.targets.
Resolves an agent name or path and returns its persona, core knowledge, tools, and rules as a structured JSON envelope for injection into the calling skill's context.
Bumps the application version following Semantic Versioning (patch/minor/major) and creates a conventional commit.
Builds a deterministic hybrid tech profile from config, file/diff auto-detection, and caller context.
Ensures feature/XXXX branch exists idempotently, locally and on origin.
(Interna) Gera o plano de arquitetura por story (arch-story-*.md) herdando as decisões de nível feature de architecture-decisions-backend.md. Produz apenas os deltas específicos da story (C4 da fatia, sequence diagrams do fluxo, data model, impacto). Não-chamável pelo usuário.
Decomposes an implementation plan into TDD tasks (Red/Green/Refactor) from test scenarios.
Generates Story artifacts from an existing Feature markdown with feature linkage and RNFs. Internal skill invoked by x-create-feature Stage 3.
Atomic read-modify-write of execution-state.json with flock and schema validation.
Loads a story file, validates predecessors, runs pre-checks, classifies scope.
Multi-agent story planning: 7 specialized agents produce task breakdown and plans.
Selects canonical knowledge assets by knowledge_kind and phase/stage via knowledge-catalog.md discovery. Technical knowledge is matched by --phase + capabilities; domain knowledge by --stage + segment/product.
Renders PR body Markdown from disk artifacts (--kind=implementation|backlog).
Selects canonical patterns/knowledge by phase and capability via catalog.md discovery. Aggregates arch_questions (from catalogs) and scoring_summary (from sub-patterns) for dynamic skill behaviour.
Java-specific developer review: dynamically loads versioned patterns based on detected Java + Spring Boot version.
Promote interactive ideation output from x-create-feature to .aikittools/ideations/.
Merges source into target locally with strategy, conflict rollback, and idempotent no-op.
Manages git worktrees for parallel task/story execution under src/worktrees/.
Git workflow: branch, atomic Conventional Commits, push, and PR creation.
Generates a Double-Loop TDD test plan with TPP-ordered acceptance and unit scenarios.
Orchestrates language-specific developer reviews in parallel based on detected technologies.
Multi-persona 4-phase TECHNICAL story refinement (post-architecture): parallel specialists + verdict.
QA review: coverage, TDD compliance, naming, fixtures, parametrized tests, AC coverage.
Observability review: tracing, metrics naming, structured logging, health, correlation.
Adds a newly discovered story to an already-refined feature; validates and proposes feature/implementation-map updates.
Full codebase review against project standards via parallel specialist subagents.
Audits dependencies for CVEs, outdated versions, and license issues per stack.
Closes a feature: validates all stories are MERGED into feature/XXXX, syncs develop, opens the final PR to develop, and generates the memory summary.
Creates Conventional Commits with Task ID and pre-commit chain (format/lint/compile).
Create a structured bug report from the RA9 template
Orchestrates Git Flow release: version bump, validation, PR, tag, back-merge.
Runs tests with coverage reporting and threshold validation.
Implements a bug story (regression-test or fix) end-to-end via BUG_GATE, ordering enforcement, context loading, and delegation to x-internal-implement-story.
Formats source code; first step of the pre-commit chain (format -> lint -> compile).
Generates ADRs from architecture-plan mini-ADRs with sequential numbering and index update.
Generates or updates CI/CD pipelines per project stack with actionlint validation.
Documentation automation v2: stack-aware generation from documentation.targets.
Hybrid changelog: narrative Highlights + Keep-a-Changelog sections from v2 features.
Implements a task via TDD Red-Green-Refactor with one atomic commit via x-commit-changes.
Orchestrates parallel story planning: 5 sibling agents (impl, tests, tasks, sec, comp).
Materializes the feature container artifact (feature-<ID>.md) from an existing Feature/spec source. Internal skill invoked by x-refine-feature Phase E.
Generates Story artifacts from an existing Feature markdown with feature linkage and RNFs. Internal skill invoked by x-create-feature Stage 3.
Parses YAML frontmatter from a patterns or knowledge file and returns structured JSON with capabilities, phases, arch_questions, scoring_config, and triggers_context_packs.
Engine compartilhada de implementação de story (Phases 0.3→3). Usada por x-implement-story e x-fix-bug.
Resolves an agent name or path and returns its persona, core knowledge, tools, and rules as a structured JSON envelope for injection into the calling skill's context.
Loads a story file, validates predecessors, runs pre-checks, classifies scope.
Renders PR body Markdown from disk artifacts (--kind=implementation|backlog).
Selects canonical knowledge assets by knowledge_kind and phase/stage via knowledge-catalog.md discovery. Technical knowledge is matched by --phase + capabilities; domain knowledge by --stage + segment/product.
Atomic read-modify-write of execution-state.json with flock and schema validation.
Validates phase transitions for orchestrators (Rule 25): pre/post/wave/final modes.
Renders _TEMPLATE-*.md from a JSON payload and writes the result atomically.
Lints source code with the appropriate linter; second step in the pre-commit chain.
Manages git worktrees for parallel task/story execution under src/worktrees/.
Merges source into target locally with strategy, conflict rollback, and idempotent no-op.
Merges a single PR via gh CLI with configurable strategy and idempotent behavior.
Generates plan-task-*.md with TDD cycles in TPP order, file-impact, and exit criteria.
Lightweight PO-centric business refinement of a feature or a story (value, problem, persona, scope, business AC, KPI). Writes a business verdict; never unblocks technical gates.
Reconciles execution-state.json telemetry against Status field of Feature/Story markdowns.
Multi-persona 4-phase TECHNICAL story refinement (post-architecture): parallel specialists + verdict.
PCI-DSS compliance review with 15-point checklist for payment-card-data changes.
Incrementally updates the service or system architecture document; never regenerative.
DevOps specialist review: Dockerfile, CI/CD, resource limits, health probes, deploy.
Observability review: tracing, metrics naming, structured logging, health, correlation.
Performance review: N+1, pools, async, pagination, cache, timeouts, circuit breakers.
QA review: coverage, TDD compliance, naming, fixtures, parametrized tests, AC coverage.
Conditional dep-policy gate: CVEs, licenses, versions, freshness; SARIF + report.
DevOps specialist review: Dockerfile, CI/CD, resource limits, health probes, deploy.
Incrementally updates the service or system architecture document; never regenerative.
Conditional dep-policy gate: CVEs, licenses, versions, freshness; SARIF + report.
Orchestrates all applicable specialist review skills in parallel; Security and QA always mandatory.
DAST gate: ZAP + Nuclei (passive/active); SARIF + Markdown; blocks on HIGH/CRITICAL.
Reviews code changes for compliance with selected security frameworks and patterns.
Tech Lead review with dynamic architecture-driven dimensions; reads all specialist and developer reports to issue GO/NO-GO.
Data-modeling specialist review: entities, aggregates, value objects, repositories, DDD.
Validates gRPC service definitions, proto3 conventions, implementation patterns, and operational readiness.
Runs integration tests that validate the complete flow from request through all application layers to response, using a real database.
Complete coding conventions: Clean Code rules (CC-01 to CC-10), SOLID principles, {{LANGUAGE}} {{LANGUAGE_VERSION}} idioms, naming patterns, constructor injection, mapper conventions, version-specific features, and approved libraries. Read before writing any code.
Executes the epic-level verification gate end-to-end (Phase 1.7 / Phase 4 carve-out of x-implement-epic) on the epic/XXXX branch HEAD: checks out the branch, runs mvn clean test + jacoco:report, parses filtered coverage against epic-level thresholds (default line >=95, branch >=90), runs a declarative DoD checklist (presence of tests, tasks DONE, CHANGELOG entry, ADR references), and emits a single-line JSON envelope {passed, failures, coverageDelta, dodChecklist}. Seventh skill in the x-internal-* convention and the fourth under internal/plan/ (after x-internal-load-story-context, x-internal-build-story-plan, and x-internal-verify-story). Isolates ~180 inline lines of integrity-gate logic from x-implement-epic.
Orchestrates multi-agent planning for all stories in an epic, respecting dependency order, with checkpoint and resume support.
Generates a comprehensive architecture plan with component diagrams, sequence diagrams, deployment topology, mini-ADRs, NFRs, and resilience/observability strategies. Use before implementation to document design decisions.
Lints source code with the appropriate linter; second step in the pre-commit chain.
Validates API contracts (OpenAPI 3.1, AsyncAPI 2.6, Protobuf 3) against their specifications. Reports structural errors, missing fields, and spec violations.
Detects and classifies file-collision risks between work units (tasks, stories, or epic phases) before parallel execution. Reads ## File Footprint blocks produced by x-plan-task / x-plan-story, applies the parallelism-heuristics knowledge pack (hard / regen / soft categories + hotspot overrides), and emits a collision matrix + serialization recommendation in Markdown (default) or JSON.
Terraform patterns: module structure, remote state, naming conventions, CI/CD workflows, drift detection, and common infrastructure modules.
Observability specialist review: validates distributed tracing, metrics naming, structured logging, health checks, correlation IDs, and alerting configuration.
Generate a Spring Boot @RestController with matching DTOs, mappers, @ControllerAdvice handler, and unit tests following hexagonal architecture.
Feature flags patterns: toggle types, lifecycle management, evaluation strategies, progressive delivery, cleanup policies, and hexagonal architecture integration.
Validates REST API endpoints for RFC 7807 error responses, pagination, URL versioning, OpenAPI documentation, status codes, and DTO patterns.
Static Application Security Testing -- scans source code for security vulnerabilities without executing the application. Produces SARIF output with OWASP mapping.
Orchestrates multi-agent planning for all stories in an epic, respecting dependency order, with checkpoint and resume support.
Detects file-collision risks between parallel work units; emits collision matrix.
Cross-cutting data modeling patterns: schema design (soft delete, temporal tables, audit trails, multi-tenant, SCD), concurrency, and test data management.
Container registry patterns: tagging strategy, immutability, retention policies, vulnerability scanning, multi-arch builds, and CI/CD integration.
Automated profiling: detect language/runtime, select appropriate profiler, execute session, generate flamegraph, identify hotspots, and suggest optimizations referencing the performance-engineering knowledge pack.
Data-modeling specialist review: entities, aggregates, value objects, repositories, DDD.
Observability specialist review: validates distributed tracing, metrics naming, structured logging, health checks, correlation IDs, and alerting configuration.
Orchestrates all applicable specialist review skills in parallel; Security and QA always mandatory.
Creates Conventional Commits with Task ID and pre-commit chain (format/lint/compile).
Runs automated smoke tests against the REST API using Newman/Postman. Supports local, container-orchestrated, and staging environments.
Analyzes source code with the appropriate linter for {{LANGUAGE}}. Second step in the pre-commit chain (RULE-007: format -> lint -> compile -> commit). Supports --fix, --changed-only, and --strict modes.
Formats source code; first step of the pre-commit chain (format -> lint -> compile).
Batch-commits planning artifacts under plans/** (or whitelisted template paths) without triggering the code pre-commit chain (format/lint/compile). Sibling of x-commit-changes for docs/markdown/json generated by planning skills.
Batch-commits planning artifacts under plans/** (or whitelisted template paths) without triggering the code pre-commit chain (format/lint/compile). Sibling of x-commit-changes for docs/markdown/json generated by planning skills.
Evaluates hardening posture (headers, TLS, CORS, cookies) with weighted SARIF scoring.
Scaffolds a Micronaut service with @Controller, DI, health, Dockerfile, and tests.
Cleans local git: prune, remove non-main worktrees, delete branches except main/develop.
Analyze telemetry NDJSON for one or more epics and produce a Markdown report with skill/phase/tool aggregates, Mermaid Gantt timeline, and optional JSON/CSV exports. Use to answer 'which phase is the bottleneck?' and 'is skill X getting slower?' questions for operator visibility.
Static Application Security Testing -- scans source code for security vulnerabilities without executing the application. Produces SARIF output with OWASP mapping.
Validates project dependencies against the declared dependency policy (denied CVEs, license whitelist, version constraints, freshness window, scope policy). Produces a structured validation report and exits with BLOCK or WARN based on D-R10/D-R11 enforcement matrix.
Implements a feature/story/task using TDD (Red-Green-Refactor) workflow. Schema-aware: v1 (legacy) runs the original Double-Loop TDD flow with story-section task extraction; v2 (task-first, EPIC-0038) reads task-TASK-XXXX-YYYY-NNN.md + plan-task-TASK-XXXX-YYYY-NNN.md, honours declared I/O contracts, respects task-implementation-map dependencies, verifies post-conditions via grep/assert, and produces a single atomic commit per task via x-commit-changes.
Parses an argv string against a declarative JSON schema with typed flags (boolean, string, integer, enum), defaults, mutually-exclusive groups, and deprecation warnings, then emits a normalized `{parsed, warnings, errors}` envelope on stdout. Replaces ~150 lines of inline argv parsing inlined inside `x-implement-epic`, `x-implement-story`, and `x-orchestrate-epic`, giving every orchestrator identical flag-validation syntax and error messages. Third skill in the x-internal-* convention (after x-internal-update-status pilot and x-internal-write-report): internal visibility, non-user-invocable, subdir scoping under internal/ops/.
Reads PR review comments and fixes actionable ones with Conventional Commits.
Git operations: branch creation, atomic commits (Conventional Commits), push, and PR creation. Use for any git workflow task including branching, committing, pushing, creating PRs, or managing version control.
Validates event schemas, producer/consumer patterns, error handling, dead letter topics, and operational readiness for event-driven architectures.
Orchestrates multi-agent planning for all stories in an epic, respecting dependency order, with checkpoint and resume support.
Release management practices: semantic versioning, version lifecycle (alpha/beta/RC/GA/LTS/EOL), release branching strategies, artifact registry management, release signing and attestation, hotfix process, rollback procedures, and release communication.
Creates a Jira Epic from a local epic markdown and syncs the Jira key back to the file.
Creates a bare git branch from a configurable base with naming validation and idempotency.
Kubernetes deployment patterns: workload types, pod specifications, resource sizing, probes, autoscaling, network policies, and security contexts.
Validates gRPC service definitions, proto3 conventions, implementation patterns, and operational readiness.
Sets up local dev: container orchestrator, database, and build tools.
Reconciles execution-state.json telemetry against Status field of Feature/Story markdowns.
Audits compliance of all project rules AND knowledge packs against source code. Launches parallel subagents (one per rule/knowledge-pack) for scanning, then aggregates into a unified report with severity classification and story suggestions.
Enhanced supply chain security audit beyond x-dependency-audit. Analyzes maintainer risk, typosquatting detection, phantom dependencies, dependency age, EPSS scoring, and SLSA assessment. Produces SARIF 2.1.0 output with weighted risk scoring.
Runs smoke tests against the TCP socket server using a standalone Java client.
Formats source code using the appropriate formatter for {{LANGUAGE}}. First step of the pre-commit chain (format -> lint -> compile -> commit). Supports --check (dry-run) and --changed-only modes.
Fonte da verdade do modelo RA9 (Rule-Aligned 9-Section) para templates de planejamento Epic/Story/Task. Define as 9 seções fixas, granularidade por nível, micro-template Decision Rationale, e mapeamento rule ↔ seção.
Reviews API gateway configuration for routing rules, authentication, rate limiting, CORS, security headers, TLS, and observability integration.
PCI-DSS v4.0 requirements mapped to code practices: 12 requirements with prohibited/correct examples and reviewer checklists.
Knowledge pack RA9 (Rule-Aligned 9-Section) for Epic/Story/Task planning templates.
Generates ADRs from architecture-plan mini-ADRs with sequential numbering and index update.
Reads PR review comments and fixes actionable ones with Conventional Commits.
Audits dependencies for CVEs, outdated versions, and license issues per stack.
Cleans local git: prune, remove non-main worktrees, delete branches except main/develop.
Analyzes telemetry NDJSON for an epic; produces Markdown report with Gantt and aggregates.
Create a structured bug report from the RA9 template
Creates a complete feature (Epic + Stories + Map) from a spec file in an isolated worktree.
Detects spec-code drift: contracts, endpoints, Gherkin scenarios vs implementation.
Stack-aware contract breaking-change detection (openapi-diff, buf, schema registry).
Detects file-collision risks between parallel work units; emits collision matrix.
Fonte da verdade do modelo RA9 (Rule-Aligned 9-Section) para templates de planejamento Epic/Story/Task. Define as 9 seções fixas, granularidade por nível, micro-template Decision Rationale, e mapeamento rule ↔ seção.
Executes the story-level verification gate (Phase 3 carve-out of x-story-implement): identifies files touched by the story via the task breakdown, runs the test suite scoped to those files, parses filtered coverage against the story-specific thresholds (default line >=95, branch >=90), performs cross-file consistency checks (constructor patterns, return-type uniformity per role), optionally runs the smoke suite, and validates every Section 7 Gherkin scenario has a matching acceptance test. Emits a single-line JSON envelope {passed, coverageDelta, failures, acCheckResults}. Sixth skill in the x-internal-* convention and the third under internal/plan/ (after x-internal-story-load-context and x-internal-story-build-plan).
Merges a single PR via gh CLI with configurable strategy (merge/squash/rebase), idempotency for already-merged PRs, pre-checks for CI and approvals in synchronous mode, GitHub native auto-merge in --auto mode, and structured error codes. Extracted from x-epic-implement Phase 1.3b to provide a testable, reusable merge primitive callable from x-pr-create --auto-merge and x-epic-implement.
Automates ADR generation from architecture plan mini-ADRs: extracts inline decisions, expands to full ADR format, assigns sequential numbering, updates the ADR index, and adds cross-references.
Generate or update CI/CD pipelines based on project stack: detect language, analyze existing workflows, generate CI/CD/release/security pipelines, validate with actionlint, support monorepo triggers.
Generate or update CI/CD pipelines based on project stack: detect language, analyze existing workflows, generate CI/CD/release/security pipelines, validate with actionlint, support monorepo triggers.
Full codebase review against all project standards. Launches parallel subagents per audit dimension (Clean Code, SOLID, Architecture, Tests, Security, Cross-file), consolidates findings into a severity-categorized report with score. Use for periodic quality validation.
Full codebase review against all project standards. Launches parallel subagents per audit dimension (Clean Code, SOLID, Architecture, Tests, Security, Cross-file), consolidates findings into a severity-categorized report with score. Use for periodic quality validation.
Analyzes source code with the appropriate linter for {{LANGUAGE}}. Second step in the pre-commit chain (RULE-007: format -> lint -> compile -> commit). Supports --fix, --changed-only, and --strict modes.
Checks project dependencies for vulnerabilities, outdated versions, and license issues. Detects build tool automatically, runs language-specific audit commands, and generates a severity-categorized report.
Checks project dependencies for vulnerabilities, outdated versions, and license issues. Detects build tool automatically, runs language-specific audit commands, and generates a severity-categorized report.
Documentation automation: detects documentation type needed (API, README, ADR, changelog) from code changes, delegates to specialized skills or generates inline. Single entry point for all documentation updates.
Documentation automation: detects documentation type needed (API, README, ADR, changelog) from code changes, delegates to specialized skills or generates inline. Single entry point for all documentation updates.
Generate an Epic document from a system specification file with cross-cutting business rules, global quality definitions (DoR/DoD), a complete story index with dependency declarations, and optional Jira integration.
Generate an Epic document from a system specification file with cross-cutting business rules, global quality definitions (DoR/DoD), a complete story index with dependency declarations, and optional Jira integration.
Complete decomposition of a system specification into an Epic, individual Story files, and an Implementation Map with dependency graph and phased execution plan. Orchestrates spec analysis, rule extraction, story identification, and implementation planning.
Generate an Implementation Map from an Epic and its Stories with dependency matrix, phase computation, critical path analysis, ASCII phase diagrams, Mermaid dependency graphs, phase summary tables, and strategic observations.
Orchestrates multi-agent planning for all stories in an epic, respecting dependency order, with checkpoint and resume support.
Creates a bare git branch (no worktree) from a configurable base with naming validation and idempotency. Single source of truth for branch creation logic consumed by orchestrators (x-internal-epic-branch-ensure, x-story-implement) and users.
Cleans local git state in one pass: fetches origin with prune, removes all non-main worktrees (any path), and deletes all local branches except main/master/develop. Destructive by default with an interactive y/N confirmation gate; supports --dry-run (preview) and --yes (non-interactive).
Creates Conventional Commits with Task ID in scope and pre-commit chain (format -> lint -> compile). Central commit point in the task-centric workflow with TDD tag support.
Creates Conventional Commits with Task ID in scope and pre-commit chain (format -> lint -> compile). Central commit point in the task-centric workflow with TDD tag support.
Manages git worktrees for parallel task and story execution. Operations: create, list, remove, cleanup, detect-context. Follows Rule 14 (Worktree Lifecycle) naming convention under .claude/worktrees/{identifier}/.
Manages git worktrees for parallel task and story execution. Operations: create, list, remove, cleanup, detect-context. Follows Rule 14 (Worktree Lifecycle) naming convention under .claude/worktrees/{identifier}/.
Builds the canonical ExecutionPlan for an epic (Phase 0/0.5 carve-out of x-epic-implement): loads epic-XXXX.md, IMPLEMENTATION-MAP.md, and every story-*.md; constructs the inter-story dependency DAG; runs Kahn's algorithm with cycle detection; optionally computes a file-overlap matrix (mode=parallel) and the critical path; then renders plans/epic-XXXX/epic-execution-plan.md via x-internal-report-write using the _TEMPLATE-EPIC-EXECUTION-PLAN.md template. Emits a stable JSON envelope on stdout for orchestrator consumption. Sixth skill in the x-internal-* convention and the third under internal/plan/ (after x-internal-story-load-context and x-internal-story-build-plan).
Executes the epic-level verification gate end-to-end (Phase 1.7 / Phase 4 carve-out of x-epic-implement) on the epic/XXXX branch HEAD: checks out the branch, runs mvn clean test + jacoco:report, parses filtered coverage against epic-level thresholds (default line >=95, branch >=90), runs a declarative DoD checklist (presence of tests, tasks DONE, CHANGELOG entry, ADR references), and emits a single-line JSON envelope {passed, failures, coverageDelta, dodChecklist}. Seventh skill in the x-internal-* convention and the fourth under internal/plan/ (after x-internal-story-load-context, x-internal-story-build-plan, and x-internal-story-verify). Isolates ~180 inline lines of integrity-gate logic from x-epic-implement.
Validates phase transitions for orchestrators under Rule 25. Four modes: --mode pre (assert predecessor phases completed before entering phase N), --mode post (assert all child tasks of phase N are completed AND all expected artifacts exist on disk), --mode wave (post-Batch-B verification of parallel wave completeness: N child TaskUpdate completed + N artifacts exist), --mode final (terminal gate composing with x-internal-epic-integrity-gate). Reads execution-state.json.taskTracking.phaseGateResults and TaskList task state; writes back the gate result. Emits a single-line JSON envelope {passed, mode, skill, phase, expectedTasks, completedTasks, missingTasks, expectedArtifacts, missingArtifacts, wallclockMs, timestamp}. Exit 0 on passed, 12 on failure, 13 on malformed args, 14 on task-resolution timeout. First skill in the x-internal-* convention authored by EPIC-0055; eighth overall (after status-update, report-write, args-normalize, story-load-context, story-build-plan, story-verify, story-resume, epic-build-plan, epic-integrity-gate, epic-branch-ensure, story-report) and the eighth under internal/plan/.
Orchestrates parallel story-planning (Phase 1 carve-out of x-story-implement): invokes x-arch-plan (Step 1A) and then dispatches 5 sibling Agent subagents in ONE assistant message for implementation plan, test plan, task breakdown, security assessment, and compliance assessment (Steps 1B-1F). Applies the Rule 13 SUBAGENT-GENERAL pattern as the canonical parallel-planning gateway. Scope-aware: SIMPLE skips 1E/1F. Returns a consolidated envelope of artifact paths to the calling orchestrator. Fifth skill in the x-internal-* convention and the second under internal/plan/ (after x-internal-story-load-context).
Loads a story file, validates predecessor dependencies against execution-state.json, executes artifact pre-checks (mtime-based staleness detection across the 7 planning artifacts), classifies scope (SIMPLE / STANDARD / COMPLEX) from task-count + Gherkin-scenario heuristics, and detects planning mode (PRE_PLANNED / HYBRID / INLINE). Replaces ~140 lines of inline Phase 0 logic previously duplicated inside x-story-implement. Fourth skill in the x-internal-* convention (after x-internal-status-update pilot, x-internal-report-write, and x-internal-args-normalize): internal visibility, non-user-invocable, read-only, subdir scoping under internal/plan/.
Generates the final consolidated story-completion report by reading plans/epic-XXXX/execution-state.json, collecting per-task status and commitSha, PR metadata (prNumber, prState), coverage delta, and review findings, then rendering the output via x-internal-report-write with _TEMPLATE-STORY-COMPLETION-REPORT.md to the caller-specified --output path. Eighth skill in the x-internal-* convention and the fifth under internal/plan/ (after x-internal-story-load-context, x-internal-story-build-plan, x-internal-story-verify, and x-internal-story-resume). Read-only against state; writes only to --output via x-internal-report-write.
Detects the resumable state of an in-flight story: reads plans/epic-XXXX/execution-state.json via x-internal-status-update --read-only, identifies the first PENDING or IN_PROGRESS task (the resume point), catalogues DONE tasks with their commitSha, extracts the last committed SHA, and flags staleness when the story file's mtime is newer than any DONE task's completion timestamp. Emits a single-line JSON envelope {resumePoint, tasksCompleted, tasksPending, lastCommitSha, staleWarnings}. Seventh skill in the x-internal-* convention and the fourth under internal/plan/ (after x-internal-story-load-context, x-internal-story-build-plan, x-internal-story-verify). Read-only by construction — never mutates state.
Executes the story-level verification gate (Phase 3 carve-out of x-story-implement): identifies files touched by the story via the task breakdown, runs the test suite scoped to those files, parses filtered coverage against the story-specific thresholds (default line >=95, branch >=90), performs cross-file consistency checks (constructor patterns, return-type uniformity per role), optionally runs the smoke suite, and validates every Section 7 Gherkin scenario has a matching acceptance test. Emits a single-line JSON envelope {passed, coverageDelta, failures, acCheckResults}. Sixth skill in the x-internal-* convention and the third under internal/plan/ (after x-internal-story-load-context and x-internal-story-build-plan).
Merges a single PR via gh CLI with configurable strategy (merge/squash/rebase), idempotency for already-merged PRs, pre-checks for CI and approvals in synchronous mode, GitHub native auto-merge in --auto mode, and structured error codes. Extracted from x-epic-implement Phase 1.3b to provide a testable, reusable merge primitive callable from x-pr-create --auto-merge and x-epic-implement.
Create a Jira Epic from an existing local epic markdown file. Read the epic file, map fields to Jira, create the issue via MCP, and sync the Jira key back to the local file.
Create a Jira Epic from an existing local epic markdown file. Read the epic file, map fields to Jira, create the issue via MCP, and sync the Jira key back to the local file.
Create Jira Stories from existing local story markdown files. Read all story files in an epic directory, map fields to Jira, create issues with parent epic link, create dependency links between stories, and sync Jira keys back to local files.
Build gate verification between parallelism groups. Compiles code, classifies errors, decides retry vs escalate, extracts outputs for next group. Used between each implementation group in Phase 2.
Decomposes an implementation plan into tasks. Primary mode: derives tasks from test scenarios (x-test-plan output) using TDD structure (RED/GREEN/REFACTOR). Fallback mode: uses Layer Task Catalog (G1-G7) when no test plan exists.
Analyzes project tech stack and recommends relevant MCP (Model Context Protocol) servers. Auto-detects language, framework, database, cache, and message broker from project config, then matches against a built-in catalog of MCP servers with installation instructions.
Analyzes project tech stack and recommends relevant MCP (Model Context Protocol) servers. Auto-detects language, framework, database, cache, and message broker from project config, then matches against a built-in catalog of MCP servers with installation instructions.
Adds or reviews distributed tracing, metrics, and structured logging using OpenTelemetry SDK with OTLP export.
Diagnoses errors, stacktraces, build failures, and unexpected behavior. Systematic approach: reproduce, locate, understand, fix, verify. Use whenever something fails: compilation errors, test failures, runtime exceptions, coverage gaps, or performance issues.
Automated OWASP Top 10 (2021) verification mapped to ASVS levels (L1/L2/L3). Checks all 10 categories (A01-A10) with per-category pass/fail, ASVS coverage percentage, score grading, SARIF 2.1.0 output, and CI integration. Delegates A06 to x-dependency-audit.
Automated OWASP Top 10 (2021) verification mapped to ASVS levels (L1/L2/L3). Checks all 10 categories (A01-A10) with per-category pass/fail, ASVS coverage percentage, score grading, SARIF 2.1.0 output, and CI integration. Delegates A06 to x-dependency-audit.
Detects and classifies file-collision risks between work units (tasks, stories, or epic phases) before parallel execution. Reads ## File Footprint blocks produced by x-task-plan / x-story-plan, applies the parallelism-heuristics knowledge pack (hard / regen / soft categories + hotspot overrides), and emits a collision matrix + serialization recommendation in Markdown (default) or JSON.
Batch-commits planning artifacts under plans/** (or whitelisted template paths) without triggering the code pre-commit chain (format/lint/compile). Sibling of x-git-commit for docs/markdown/json generated by planning skills.
Reads PR review comments and fixes actionable ones automatically. Detects PR from argument or branch, classifies comments (actionable/suggestion/question/praise), implements fixes, and commits with proper conventional commit messages.
Reads PR review comments and fixes actionable ones automatically. Detects PR from argument or branch, classifies comments (actionable/suggestion/question/praise), implements fixes, and commits with proper conventional commit messages.
Merge-train automation: discovers, validates, and merges a sequence of PRs into develop in deterministic order. Supports --prs, --epic, and --pattern discovery modes with pre-merge validation and dry-run auditing.
Polls a PR's CI checks and Copilot review status, blocking until checks complete or timeout. Returns one of 8 stable exit codes (SUCCESS=0, CI_PENDING_PROCEED=10, CI_FAILED=20, TIMEOUT=30, PR_ALREADY_MERGED=40, NO_CI_CONFIGURED=50, PR_CLOSED=60, PR_NOT_FOUND=70). Writes a versioned state-file for session resume.
Polls a PR's CI checks and Copilot review status, blocking until checks complete or timeout. Returns one of 8 stable exit codes (SUCCESS=0, CI_PENDING_PROCEED=10, CI_FAILED=20, TIMEOUT=30, PR_ALREADY_MERGED=40, NO_CI_CONFIGURED=50, PR_CLOSED=60, PR_NOT_FOUND=70). Writes a versioned state-file for session resume.
Orchestrates complete release flow using Git Flow release branches with approval gate, PR-flow (gh CLI) and deep validation: version bump (auto-detect or explicit), release branch creation from develop, deep validation (coverage, golden files, version consistency), version file updates, changelog generation, release commit, release PR via gh (optionally reviewed by x-review-pr), human approval gate with persistent state file, tag on main after merged PR, back-merge PR to develop with conflict detection, and cleanup. Supports hotfix releases from main, dry-run mode, resume via --continue-after-merge, in-session pause via --interactive, GPG-signed tags, skip-review opt-out, and custom state file path.
Generates CHANGELOG.md from Conventional Commits history. Parses git log, groups by commit type, maps to Keep a Changelog sections (Added, Changed, Fixed, etc.), and performs incremental updates preserving existing entries.
Orchestrates complete release flow using Git Flow release branches with approval gate, PR-flow (gh CLI) and deep validation: version bump (auto-detect or explicit), release branch creation from develop, deep validation (coverage, golden files, version consistency), version file updates, changelog generation, release commit, release PR via gh (optionally reviewed by x-review-pr), human approval gate with persistent state file, tag on main after merged PR, back-merge PR to develop with conflict detection, and cleanup. Supports hotfix releases from main, dry-run mode, resume via --continue-after-merge, in-session pause via --interactive, GPG-signed tags, skip-review opt-out, and custom state file path.
Parallel code review with specialist engineers (Security, QA, Performance, Database, Observability, DevOps, API, Event). Invokes individual review skills in parallel via Skill tool, then consolidates into a scored report. Use for pre-PR quality validation.
Validates REST API endpoints for RFC 7807 error responses, pagination, URL versioning, OpenAPI documentation, status codes, and DTO patterns.
PCI-DSS compliance review with 25-point checklist for code changes involving payment card data. Produces per-point PASS/FAIL report with remediation.
Data modeling specialist review: validates entity design, aggregate boundaries, value objects, repository patterns, domain event persistence, and DDD tactical patterns.
Validates event schemas, producer/consumer patterns, error handling, dead letter topics, and operational readiness for event-driven architectures.
Parallel code review with specialist engineers (Security, QA, Performance, Database, Observability, DevOps, API, Event). Invokes individual review skills in parallel via Skill tool, then consolidates into a scored report. Use for pre-PR quality validation.
Observability specialist review: validates distributed tracing, metrics naming, structured logging, health checks, correlation IDs, and alerting configuration.
Performance specialist review: validates N+1 queries, connection pools, async patterns, pagination, caching, timeouts, circuit breakers, and resource cleanup.
Performance specialist review: validates N+1 queries, connection pools, async patterns, pagination, caching, timeouts, circuit breakers, and resource cleanup.
QA specialist review: validates test coverage, TDD compliance, test naming, fixtures, parametrized tests, and acceptance criteria coverage.
Reviews code changes for compliance with selected security frameworks. Verifies sensitive data handling, audit trails, and access control patterns.
Evaluate runtime protection controls: rate limiting, WAF rules, bot protection, DDoS mitigation, account lockout, brute force protection, CSP enforcement, and permissions policy. Produce SARIF 2.1.0 output with ASVS compliance mapping and scored Markdown report.
Scans Docker images for CVEs and Dockerfile best practices violations. Uses Trivy, Grype, or Snyk Container. Produces SARIF output with scoring.
Aggregates results from all security scanning skills into a unified posture view with score 0-100, trend tracking, OWASP risk heatmap, per-dimension breakdown, and remediation priority queue. Never executes scans — reads existing results only (RULE-011).
Dynamic Application Security Testing -- tests the running application for XSS, injection, misconfiguration, and information disclosure using OWASP ZAP or Nuclei.
Multi-phase penetration test orchestrator: reconnaissance, vulnerability scanning, exploitation validation, and consolidated reporting with environment-based restrictions.
Generate CI/CD pipeline configurations with conditional security stages based on SecurityConfig flags. Support GitHub Actions, GitLab CI, and Azure DevOps with minimal and full stage modes, configurable severity thresholds, and SARIF artifact upload.
Generate CI/CD pipeline configurations with conditional security stages based on SecurityConfig flags. Support GitHub Actions, GitLab CI, and Azure DevOps with minimal and full stage modes, configurable severity thresholds, and SARIF artifact upload.
Static Application Security Testing -- scans source code for security vulnerabilities without executing the application. Produces SARIF output with OWASP mapping.
Integrates with SonarQube/SonarCloud for security hotspot tracking, quality gate enforcement, and SARIF output from findings.
Detects spec-code drift by comparing story data contracts, endpoints, and Gherkin scenarios against implemented code. Supports standalone mode (full report) and inline mode (compact output for TDD loop integration in x-story-implement Phase 2).
Detects spec-code drift by comparing story data contracts, endpoints, and Gherkin scenarios against implemented code. Supports standalone mode (full report) and inline mode (compact output for TDD loop integration in x-story-implement Phase 2).
Multi-agent story planning: launches 5 specialized agents (Architect, QA, Security, Tech Lead, Product Owner) in parallel to produce a consolidated task breakdown, individual task plans, planning report, and DoR validation. Schema-aware: v1 (legacy) runs the original 6-phase flow; v2 (task-first, EPIC-0038) adds Phases 4a-4c that emit task-TASK-NNN.md + plan-task-TASK-NNN.md per task and a task-implementation-map-STORY-*.md, wiring every task through x-task-plan in parallel.
Enhanced supply chain security audit beyond x-dependency-audit. Analyzes maintainer risk, typosquatting detection, phantom dependencies, dependency age, EPSS scoring, and SLSA assessment. Produces SARIF 2.1.0 output with weighted risk scoring.
Analyze telemetry NDJSON for one or more epics and produce a Markdown report with skill/phase/tool aggregates, Mermaid Gantt timeline, and optional JSON/CSV exports. Use to answer 'which phase is the bottleneck?' and 'is skill X getting slower?' questions for operator visibility.
Detect cross-epic P95 regressions (>= threshold %) and rank top-10 slowest skills from the global telemetry index. Single-responsibility partner of /x-telemetry-analyze focused on trend detection, not point-in-time reporting. Use to answer 'is skill X getting slower over the last N epics?' with evidence.
Detect cross-epic P95 regressions (>= threshold %) and rank top-10 slowest skills from the global telemetry index. Single-responsibility partner of /x-telemetry-analyze focused on trend detection, not point-in-time reporting. Use to answer 'is skill X getting slower over the last N epics?' with evidence.
Runs consumer-driven contract tests (Pact, Spring Cloud Contract) to verify API compatibility between services.
Runs consumer-driven contract tests (Pact, Spring Cloud Contract) to verify API compatibility between services.
Validates API contracts (OpenAPI 3.1, AsyncAPI 2.6, Protobuf 3) against their specifications. Reports structural errors, missing fields, and spec violations.
Runs performance tests to validate latency SLAs, throughput targets, and resource stability under load. Supports baseline, normal, peak, and sustained scenarios.
Runs performance tests to validate latency SLAs, throughput targets, and resource stability under load. Supports baseline, normal, peak, and sustained scenarios.
Runs automated smoke tests against the TCP socket server using a standalone Java client with message framing and protocol validation.
Executes systematic Red-Green-Refactor TDD cycles for a task. Reads the task plan generated by x-task-plan, runs each cycle in TPP order, validates RED/GREEN/REFACTOR phases, delegates atomic commits to x-git-commit with TDD tags, and supports resume and dry-run.
Generate threat models using STRIDE analysis: identify components, map data flows, analyze threats per category, classify severity, suggest mitigations, and produce threat model document.
Generate threat models using STRIDE analysis: identify components, map data flows, analyze threats per category, classify severity, suggest mitigations, and produce threat model document.
Git operations: branch creation, atomic commits (Conventional Commits), push, and PR creation. Use for any git workflow task including branching, committing, pushing, creating PRs, or managing version control.
Batch-commits planning artifacts under plans/** (or whitelisted template paths) without triggering the code pre-commit chain (format/lint/compile). Sibling of x-git-commit for docs/markdown/json generated by planning skills.
Canonical catalog of file-collision heuristics for parallel task, story, and epic execution. Defines the File Footprint block format, the three Conflict Categories (hard / regen / soft), a hotspot list of high-contention paths, and the degrade-with-warning policy. Consumed by x-task-plan, x-story-plan, x-epic-map, x-parallel-eval, and x-dev-*-implement.
Database specialist review: validates schema design, migration safety, query optimization, connection management, transaction boundaries, and data integrity patterns.
Automated profiling: detect language/runtime, select appropriate profiler, execute session, generate flamegraph, identify hotspots, and suggest optimizations referencing the performance-engineering knowledge pack.
Evaluate runtime protection controls: rate limiting, WAF rules, bot protection, DDoS mitigation, account lockout, brute force protection, CSP enforcement, and permissions policy. Produce SARIF 2.1.0 output with ASVS compliance mapping and scored Markdown report.
Complete coding conventions: Clean Code rules (CC-01 to CC-10), SOLID principles, {{LANGUAGE}} {{LANGUAGE_VERSION}} idioms, naming patterns, constructor injection, mapper conventions, version-specific features, and approved libraries. Read before writing any code.
Scans Kubernetes manifests, Terraform modules, Helm charts, and Docker Compose files for misconfigurations against CIS benchmarks.
Validates API contracts (OpenAPI 3.1, AsyncAPI 2.6, Protobuf 3) against their specifications. Reports structural errors, missing fields, and spec violations.
API design principles: {{LANGUAGE}}-specific patterns for REST/gRPC/GraphQL. URL structure, status codes, RFC 7807 errors, pagination, content negotiation, validation, request/response shaping, versioning strategies, and protocol conventions.
CI/CD pipeline patterns: build stages, test parallelization, security scanning, artifact management, environment promotion, approval gates, caching strategies, and language-specific pipeline configurations.
FinOps practices: resource rightsizing, cost allocation, spot instances, reserved capacity, cost alerting, and cloud-specific cost optimization tools
SRE practices: error budgets, toil reduction, on-call practices, capacity planning, incident management process, and change management. Covers SLO-based release gates, burn rate calculation, automation prioritization, rotation patterns, load testing methodology, blameless postmortems, and canary analysis.
Resilience patterns: circuit breaker, rate limiting, bulkhead isolation, timeout control, retry with exponential backoff + jitter, fallback/graceful degradation, backpressure, and resilience metrics.
Release management practices: semantic versioning, version lifecycle (alpha/beta/RC/GA/LTS/EOL), release branching strategies, artifact registry management, release signing and attestation, hotfix process, rollback procedures, and release communication.
Kustomize patterns: directory structure, patches, components, secret management, generators, and patch types for environment management.
Protocol conventions: REST (OpenAPI 3.1), gRPC (Proto3), GraphQL, WebSocket, and event-driven messaging. URL structure, versioning, error handling per protocol, schema design, and integration patterns.
Reference code templates for each hexagonal architecture layer. Provides consistent patterns for domain model, ports, DTOs, mappers, entities, repositories, use cases, REST resources, exception mappers, migrations, and configuration. Uses {{LANGUAGE}}, {{FRAMEWORK}} placeholders.
Crossplane patterns: CompositeResourceDefinitions, Compositions, Claims, Provider configuration, and comparison with Terraform.
Complete security reference: OWASP Top 10, security headers, secrets management, input validation, cryptography (TLS, hashing, key management), and pentest readiness checklist. Read during security reviews or when implementing security-sensitive features.
Skill canônica de planejamento de arquitetura (níveis product/capability/feature): conduz a entrevista de decisões, gera tech-context.config.json e diagramas C4 (Context/Container/Component) em Mermaid ou PlantUML.
Feature flags patterns: toggle types, lifecycle management, evaluation strategies, progressive delivery, cleanup policies, and hexagonal architecture integration.
FinOps practices: resource rightsizing, cost allocation, spot instances, reserved capacity, cost alerting, and cloud-specific cost optimization tools
Data management lifecycle patterns: zero-downtime migrations, expand/contract pattern, schema versioning, data governance, backup/restore strategies, partitioning, CDC, and data quality validation for {{DATABASE_TYPE}} with {{MIGRATION_TOOL}}.
DDD Strategic Design: bounded context identification, context map with 6 integration patterns, Anti-Corruption Layer template, and /x-ddd-context-map skill.
Database conventions for {{DB_TYPE}} + {{CACHE_TYPE}}: schema design, migrations, indexing, query optimization, caching patterns, and connection pool management.
Resilience patterns: circuit breaker, rate limiting, bulkhead isolation, timeout control, retry with exponential backoff + jitter, fallback/graceful degradation, backpressure, and resilience metrics.
Generates the final consolidated story-completion report by reading plans/epic-XXXX/execution-state.json, collecting per-task status and commitSha, PR metadata (prNumber, prState), coverage delta, and review findings, then rendering the output via x-internal-report-write with _TEMPLATE-STORY-COMPLETION-REPORT.md to the caller-specified --output path. Eighth skill in the x-internal-* convention and the fifth under internal/plan/ (after x-internal-story-load-context, x-internal-story-build-plan, x-internal-story-verify, and x-internal-story-resume). Read-only against state; writes only to --output via x-internal-report-write.
Orchestrates parallel story-planning (Phase 1 carve-out of x-story-implement): invokes x-arch-plan (Step 1A) and then dispatches 5 sibling Agent subagents in ONE assistant message for implementation plan, test plan, task breakdown, security assessment, and compliance assessment (Steps 1B-1F). Applies the Rule 13 SUBAGENT-GENERAL pattern as the canonical parallel-planning gateway. Scope-aware: SIMPLE skips 1E/1F. Returns a consolidated envelope of artifact paths to the calling orchestrator. Fifth skill in the x-internal-* convention and the second under internal/plan/ (after x-internal-story-load-context).
Complete testing reference: testing philosophy, 8 test categories, coverage thresholds, fixture patterns, data uniqueness, async handling, database strategy, and {{LANGUAGE}}-specific test frameworks. Read before writing tests.
Merges a source branch into a target branch locally with configurable strategy (merge/squash/rebase), automatic conflict detection + rollback, and idempotent no-op when target already contains source HEAD. Centralizes the ~120 lines of inline Bash previously in x-epic-implement Phase 1.4e auto-rebase.
Story decomposition and planning: layer-by-layer decomposition (foundation, core domain, extensions, compositions, cross-cutting), story self-containment (data contracts, acceptance criteria), dependency DAG, sizing rules, and phase computation.
Skill canônica de planejamento de arquitetura (níveis product/capability/feature): conduz a entrevista de decisões, gera tech-context.config.json e diagramas C4 (Context/Container/Component) em Mermaid ou PlantUML.
Reviews API gateway configuration for routing rules, authentication, rate limiting, CORS, security headers, TLS, and observability integration.
Guides incident response with severity-based checklists, communication templates, and postmortem triggers. Interactive guide for SEV1-SEV4 incidents covering classification, response coordination, and action item tracking.
Canonical catalog of file-collision heuristics for parallel task, story, and epic execution. Defines the File Footprint block format, the three Conflict Categories (hard / regen / soft), a hotspot list of high-contention paths, and the degrade-with-warning policy. Consumed by x-task-plan, x-story-plan, x-epic-map, x-parallel-eval, and x-dev-*-implement.
OWASP ASVS 4.0.3 verification standard with L1/L2/L3 levels, cross-reference tables, and verification items for all 14 chapters (V1-V14).
Reference code templates for each hexagonal architecture layer. Provides consistent patterns for domain model, ports, DTOs, mappers, entities, repositories, use cases, REST resources, exception mappers, migrations, and configuration. Uses {{LANGUAGE}}, {{FRAMEWORK}} placeholders.
PCI-DSS v4.0 requirements mapped to code practices: 12 requirements with prohibited/correct examples and reviewer checklists.
Infrastructure patterns: Docker multi-stage builds, Kubernetes manifests (cloud-agnostic), security context, 12-Factor App principles, graceful shutdown, resource management, and cloud-native design.
Helm chart patterns: chart structure, values templates, multi-environment configuration, dependencies, testing, GitOps integration, and Helmfile.
Terraform patterns: module structure, remote state, naming conventions, CI/CD workflows, drift detection, and common infrastructure modules.
Container registry patterns: tagging strategy, immutability, retention policies, vulnerability scanning, multi-arch builds, and CI/CD integration.
Crossplane patterns: CompositeResourceDefinitions, Compositions, Claims, Provider configuration, and comparison with Terraform.
Single source of truth for the `epic/<ID>` branch convention (RULE-001). Invoked by every epic entry-point (x-epic-create, x-epic-decompose, x-epic-orchestrate, x-epic-implement, x-epic-map) at step 0; ensures the branch exists idempotently both locally and on origin. When the branch is absent, delegates creation to `x-git-branch`; when present locally but not on origin, emits a complementary `git push`. Seventh skill in the x-internal-* convention and the first under internal/git/ (after x-internal-status-update / x-internal-report-write / x-internal-args-normalize at internal/ops/ and x-internal-story-load-context / x-internal-story-build-plan / x-internal-story-verify at internal/plan/).
Merges a source branch into a target branch locally with configurable strategy (merge/squash/rebase), automatic conflict detection + rollback, and idempotent no-op when target already contains source HEAD. Centralizes the ~120 lines of inline Bash previously in x-epic-implement Phase 1.4e auto-rebase.
Detects and classifies file-collision risks between work units (tasks, stories, or epic phases) before parallel execution. Reads ## File Footprint blocks produced by x-task-plan / x-story-plan, applies the parallelism-heuristics knowledge pack (hard / regen / soft categories + hotspot overrides), and emits a collision matrix + serialization recommendation in Markdown (default) or JSON.
Full architecture reference: {{ARCHITECTURE}} principles, package structure, dependency rules, thread-safety, mapper patterns, persistence rules, and architecture variants. Read before designing or implementing features.
Validate RNF no-relax markers and justification gate for override inheritance
Validate RNF no-relax markers and justification gate for override inheritance
Detects and classifies file-collision risks between work units (tasks, stories, or epic phases) before parallel execution. Reads ## File Footprint blocks produced by x-task-plan / x-story-plan, applies the parallelism-heuristics knowledge pack (hard / regen / soft categories + hotspot overrides), and emits a collision matrix + serialization recommendation in Markdown (default) or JSON.
Complete decomposition of a system specification into an Epic, individual Story files, and an Implementation Map with dependency graph and phased execution plan. Orchestrates spec analysis, rule extraction, story identification, and implementation planning.
Git workflow: branch, atomic Conventional Commits, push, and PR creation.
Loads a story file, validates predecessor dependencies against execution-state.json, executes artifact pre-checks (mtime-based staleness detection across the 7 planning artifacts), classifies scope (SIMPLE / STANDARD / COMPLEX) from task-count + Gherkin-scenario heuristics, and detects planning mode (PRE_PLANNED / HYBRID / INLINE). Replaces ~140 lines of inline Phase 0 logic previously duplicated inside x-story-implement. Fourth skill in the x-internal-* convention (after x-internal-status-update pilot, x-internal-report-write, and x-internal-args-normalize): internal visibility, non-user-invocable, read-only, subdir scoping under internal/plan/.
Database conventions for {{DB_TYPE}} + {{CACHE_TYPE}}: schema design, migrations, indexing, query optimization, caching patterns, and connection pool management.
Dockerfile patterns: multi-stage builds, security hardening, .dockerignore templates, layer optimization, health checks, and OCI labels per language.
Protocol conventions: REST (OpenAPI 3.1), gRPC (Proto3), GraphQL, WebSocket, and event-driven messaging. URL structure, versioning, error handling per protocol, schema design, and integration patterns.
Performance engineering patterns: profiling, benchmarking, optimization, regression detection, and memory management for {{LANGUAGE}} {{FRAMEWORK}} projects using {{BUILD_TOOL}}.
Spring Boot patterns: DI, @ConfigurationProperties, Spring Data JPA, @RestController, @ControllerAdvice, Spring AOT, and health checks.
Single source of truth for the `epic/<ID>` branch convention (RULE-001). Invoked by every epic entry-point (x-epic-create, x-epic-decompose, x-epic-orchestrate, x-epic-implement, x-epic-map) at step 0; ensures the branch exists idempotently both locally and on origin. When the branch is absent, delegates creation to `x-git-branch`; when present locally but not on origin, emits a complementary `git push`. Seventh skill in the x-internal-* convention and the first under internal/git/ (after x-internal-status-update / x-internal-report-write / x-internal-args-normalize at internal/ops/ and x-internal-story-load-context / x-internal-story-build-plan / x-internal-story-verify at internal/plan/).
Reconciles execution-state.json (telemetry) against the **Status:** field of Epic / Story markdown artifacts. Default mode (diagnose) is read-only and prints a divergence table. Opt-in --apply rewrites the markdowns atomically via StatusFieldParser and commits via x-git-commit. Respects Rule 19 (legacy v1 epics skip silently) and Rule 22 (markdown is SoT; state.json is telemetry). Use for manual recovery of legacy epics whose markdown status drifted from execution checkpoints.
Multi-persona 4-phase strategic feature refinement: parallel specialists + verdict. On approval, auto-creates the feature container.
Runs integration tests that validate the complete flow from request through all application layers to response, using a real database.
Generate or update CI/CD pipelines based on project stack: detect language, analyze existing workflows, generate CI/CD/release/security pipelines, validate with actionlint, support monorepo triggers.
Loads a story file, validates predecessor dependencies against execution-state.json, executes artifact pre-checks (mtime-based staleness detection across the 7 planning artifacts), classifies scope (SIMPLE / STANDARD / COMPLEX) from task-count + Gherkin-scenario heuristics, and detects planning mode (PRE_PLANNED / HYBRID / INLINE). Replaces ~140 lines of inline Phase 0 logic previously duplicated inside x-implement-story. Fourth skill in the x-internal-* convention (after x-internal-update-status pilot, x-internal-write-report, and x-internal-normalize-args): internal visibility, non-user-invocable, read-only, subdir scoping under internal/plan/.
Analyzes project tech stack and recommends relevant MCP (Model Context Protocol) servers. Auto-detects language, framework, database, cache, and message broker from project config, then matches against a built-in catalog of MCP servers with installation instructions.
Generates a Picocli @Command with subcommands, options, converters, and unit tests.
Formats source code using the appropriate formatter for {{LANGUAGE}}. First step of the pre-commit chain (format -> lint -> compile -> commit). Supports --check (dry-run) and --changed-only modes.
Adds or reviews distributed tracing, metrics, and structured logging using OpenTelemetry SDK with OTLP export.
Validates REST API endpoints for RFC 7807 error responses, pagination, URL versioning, OpenAPI documentation, status codes, and DTO patterns.
Data modeling specialist review: validates entity design, aggregate boundaries, value objects, repository patterns, domain event persistence, and DDD tactical patterns.
Database specialist review: validates schema design, migration safety, query optimization, connection management, transaction boundaries, and data integrity patterns.
DevOps specialist review: validates Dockerfile, container security, CI/CD pipeline, resource limits, health probes, graceful shutdown, and deployment configuration.
Reviews code changes for compliance with selected security frameworks. Verifies sensitive data handling, audit trails, and access control patterns.
Validates GraphQL schema design, resolver implementation, security patterns, and observability for compliance with best practices.
PCI-DSS compliance review with 25-point checklist for code changes involving payment card data. Produces per-point PASS/FAIL report with remediation.
DevOps specialist review: validates Dockerfile, container security, CI/CD pipeline, resource limits, health probes, graceful shutdown, and deployment configuration.
Scans code and git history for leaked credentials, API keys, tokens, and secrets. Produces SARIF output with scoring and baseline support.
Runs integration tests that validate the complete flow from request through all application layers to response, using a real database.
Runs automated smoke tests against the REST API using Newman/Postman. Supports local, container-orchestrated, and staging environments.
Stack-aware mutation testing skill: detects language/framework, dispatches to PIT (Java/Maven or Gradle), Stryker (JS/TS), mutmut (Python), or go-mutesting (Go), enforces quality.mutation.threshold (default 80%) and runtime-cap-min, and blocks merge on score below threshold.
Single source of truth for the `epic/<ID>` branch convention (RULE-001). Invoked by every epic entry-point (x-epic-create, x-epic-decompose, x-orchestrate-epic, x-implement-epic, x-epic-map) at step 0; ensures the branch exists idempotently both locally and on origin. When the branch is absent, delegates creation to `x-create-git-branch`; when present locally but not on origin, emits a complementary `git push`. Seventh skill in the x-internal-* convention and the first under internal/git/ (after x-internal-update-status / x-internal-write-report / x-internal-normalize-args at internal/ops/ and x-internal-load-story-context / x-internal-build-story-plan / x-internal-verify-story at internal/plan/).
Runs automated smoke tests against the TCP socket server using a standalone Java client with message framing and protocol validation.
Validates API contracts (OpenAPI 3.1, AsyncAPI 2.6, Protobuf 3) against their specifications. Reports structural errors, missing fields, and spec violations.
Analyzes source code with the appropriate linter for {{LANGUAGE}}. Second step in the pre-commit chain (RULE-007: format -> lint -> compile -> commit). Supports --fix, --changed-only, and --strict modes.
Scaffold a Micronaut service with @Controller, @Singleton DI, health indicators, compile-time DI, Dockerfile, and integration tests.
Generate or update CI/CD pipelines based on project stack: detect language, analyze existing workflows, generate CI/CD/release/security pipelines, validate with actionlint, support monorepo triggers.
Creates a bare git branch (no worktree) from a configurable base with naming validation and idempotency. Single source of truth for branch creation logic consumed by orchestrators (x-internal-ensure-epic-branch, x-implement-story) and users.
Renders _TEMPLATE-*.md templates by substituting {{KEY}} placeholders (simple and nested, dot-path) and resolving {{#each}} loops against a structured JSON data payload, then writes the result atomically to an output path. Supports an --append mode with per-section deduplication keyed by `## ID: <value>` markers. Centralises phase reports, epic execution reports and planning reports so orchestrators (x-implement-epic, x-orchestrate-epic, x-implement-story) stop duplicating `Read template + inline Edit/Write` logic. Second skill in the x-internal-* convention (after x-internal-update-status pilot): internal visibility, non-user-invocable, subdir scoping under internal/ops/.
Builds the canonical ExecutionPlan for an epic (Phase 0/0.5 carve-out of x-implement-epic): loads epic-XXXX.md, IMPLEMENTATION-MAP.md, and every story-*.md; constructs the inter-story dependency DAG; runs Kahn's algorithm with cycle detection; optionally computes a file-overlap matrix (mode=parallel) and the critical path; then renders ai/epics/epic-XXXX/epic-execution-plan.md via x-internal-write-report using the _TEMPLATE-EPIC-EXECUTION-PLAN.md template. Emits a stable JSON envelope on stdout for orchestrator consumption. Sixth skill in the x-internal-* convention and the third under internal/plan/ (after x-internal-load-story-context and x-internal-build-story-plan).
Generate an Epic document from spec analysis: cross-cutting rules, story index, DoR/DoD, optional Jira. Invoked only by x-create-feature (Phase 2). Not user-invocable.
Generates the final consolidated story-completion report by reading ai/epics/epic-XXXX/execution-state.json, collecting per-task status and commitSha, PR metadata (prNumber, prState), coverage delta, and review findings, then rendering the output via x-internal-write-report with _TEMPLATE-STORY-COMPLETION-REPORT.md to the caller-specified --output path. Eighth skill in the x-internal-* convention and the fifth under internal/plan/ (after x-internal-load-story-context, x-internal-build-story-plan, x-internal-verify-story, and x-internal-resume-story). Read-only against state; writes only to --output via x-internal-write-report.
Migrates artifact frontmatter from v2 to v3.0 (requires-capabilities) using path heuristics, keyword scan, and AI fallback
Renders PR body Markdown from existing disk artifacts using one of two templates. --kind=implementation: reads review/verify/telemetry artifacts for a story and emits a structured implementation PR body. --kind=backlog: reads epic/map artifacts and emits a scaffolding PR body. Fail-open: absent artifacts produce human-readable placeholders, never audit-sentinel strings. Internal — invoked only by x-pr-create and x-feature-create.
Renders PR body Markdown from existing disk artifacts using one of two templates. --kind=implementation: reads review/verify/telemetry artifacts for a story and emits a structured implementation PR body. --kind=backlog: reads epic/map artifacts and emits a scaffolding PR body. Fail-open: absent artifacts produce human-readable placeholders, never audit-sentinel strings. Internal — invoked only by x-create-pr and x-create-feature.
Create a Jira Epic from an existing local epic markdown file. Read the epic file, map fields to Jira, create the issue via MCP, and sync the Jira key back to the local file.
Queries ai/memory/ for decisions, patterns, and anti-patterns across past epics
Create a Jira Epic from an existing local epic markdown file. Read the epic file, map fields to Jira, create the issue via MCP, and sync the Jira key back to the local file.
Orchestrates complete release flow using Git Flow release branches with approval gate, PR-flow (gh CLI) and deep validation: version bump (auto-detect or explicit), release branch creation from develop, deep validation (coverage, golden files, version consistency), version file updates, changelog generation, release commit, release PR via gh (optionally reviewed by x-review-pr), human approval gate with persistent state file, tag on main after merged PR, back-merge PR to develop with conflict detection, and cleanup. Supports hotfix releases from main, dry-run mode, resume via --continue-after-merge, in-session pause via --interactive, GPG-signed tags, skip-review opt-out, and custom state file path.
Create a complete feature (Epic + N Stories + Implementation Map) from a spec file, with a worktree-isolated docs/ branch, consolidated commit, and auto-merged PR into epic/XXXX.
Generate threat models using STRIDE analysis: identify components, map data flows, analyze threats per category, classify severity, suggest mitigations, and produce threat model document.
Generates a comprehensive architecture plan with component diagrams, sequence diagrams, deployment topology, mini-ADRs, NFRs, and resilience/observability strategies. Use before implementation to document design decisions.
Sets up local dev: container orchestrator, database, and build tools.
Adds or reviews distributed tracing, metrics, and structured logging using OpenTelemetry SDK with OTLP export.
Observability specialist review: validates distributed tracing, metrics naming, structured logging, health checks, correlation IDs, and alerting configuration.
Validates GraphQL schema design, resolver implementation, security patterns, and observability for compliance with best practices.
Dynamic Application Security Testing -- tests the running application for XSS, injection, misconfiguration, and information disclosure using OWASP ZAP or Nuclei.
Multi-phase penetration test orchestrator: reconnaissance, vulnerability scanning, exploitation validation, and consolidated reporting with environment-based restrictions.
Scaffold a Helidon SE or MP service with routing, health checks, config, Dockerfile, and integration tests.
Runs performance tests to validate latency SLAs, throughput targets, and resource stability under load. Supports baseline, normal, peak, and sustained scenarios.
Scaffold a Micronaut service with @Controller, @Singleton DI, health indicators, compile-time DI, Dockerfile, and integration tests.
Validate and configure local development environment: detect stack, check prerequisites, verify versions, validate IDE config, test database connectivity, run initial build, and report status with fix suggestions.
Creates a bare git branch (no worktree) from a configurable base with naming validation and idempotency. Single source of truth for branch creation logic consumed by orchestrators (x-internal-ensure-epic-branch, x-implement-story) and users.
Merges a source branch into a target branch locally with configurable strategy (merge/squash/rebase), automatic conflict detection + rollback, and idempotent no-op when target already contains source HEAD. Centralizes the ~120 lines of inline Bash previously in x-implement-epic Phase 1.4e auto-rebase.
Classifies the git working tree state (CLEAN/DIRTY/DIVERGENT/AMBIGUOUS) and returns a stable exit code. Invoked by orchestrators before starting story/task execution to detect dirty or divergent states early. Exit 0 = CLEAN or DIRTY; exit 15 (WORKTREE_AMBIGUOUS) = DIVERGENT or AMBIGUOUS (unless --allow-dirty).
Single source of truth for the `epic/<ID>` branch convention (RULE-001). Invoked by every epic entry-point (x-epic-create, x-epic-decompose, x-orchestrate-epic, x-implement-epic, x-epic-map) at step 0; ensures the branch exists idempotently both locally and on origin. When the branch is absent, delegates creation to `x-create-git-branch`; when present locally but not on origin, emits a complementary `git push`. Seventh skill in the x-internal-* convention and the first under internal/git/ (after x-internal-update-status / x-internal-write-report / x-internal-normalize-args at internal/ops/ and x-internal-load-story-context / x-internal-build-story-plan / x-internal-verify-story at internal/plan/).
Renders PR body Markdown from existing disk artifacts using one of two templates. --kind=implementation: reads review/verify/telemetry artifacts for a story and emits a structured implementation PR body. --kind=backlog: reads epic/map artifacts and emits a scaffolding PR body. Fail-open: absent artifacts produce human-readable placeholders, never audit-sentinel strings. Internal — invoked only by x-create-pr and x-create-feature.
Queries ai/memory/ for decisions, patterns, and anti-patterns across past epics
Create a complete feature (Epic + N Stories + Implementation Map) from a spec file, with a worktree-isolated docs/ branch, consolidated commit, and auto-merged PR into epic/XXXX.
Automates ADR generation from architecture plan mini-ADRs: extracts inline decisions, expands to full ADR format, assigns sequential numbering, updates the ADR index, and adds cross-references.
Generates a detailed per-task implementation plan (plan-task-TASK-XXXX-YYYY-NNN.md) with TDD cycles in TPP order, file impact analysis by architecture layer, security checklist by task type, and exit criteria. Two invocation modes: task-file-first (--task-file) consumes a standalone task-TASK-XXXX-YYYY-NNN.md contract (EPIC-0038); story-scoped (STORY-ID --task TASK-ID) reads the task from story Section 8 (legacy). Invocable standalone OR via x-plan-story (future).
Incrementally updates the service architecture document with changes from architecture plans. Adds new components, integrations, flows, and ADR references without rewriting existing content. Use after implementation to keep architecture documentation current.
Full architecture reference: {{ARCHITECTURE}} principles, package structure, dependency rules, thread-safety, mapper patterns, persistence rules, and architecture variants. Read before designing or implementing features.
Multi-phase penetration test orchestrator: reconnaissance, vulnerability scanning, exploitation validation, and consolidated reporting with environment-based restrictions.
Reviews API gateway configuration for routing rules, authentication, rate limiting, CORS, security headers, TLS, and observability integration.
Dynamic Application Security Testing -- tests the running application for XSS, injection, misconfiguration, and information disclosure using OWASP ZAP or Nuclei.
Scans Kubernetes manifests, Terraform modules, Helm charts, and Docker Compose files for misconfigurations against CIS benchmarks.
Regression-shell gate: reads QualityConfig.regression, runs curated scenario scripts against the target (self mode: generator output; service mode: client services), compares against baseline, and blocks merge on unexpected diffs.
PCI-DSS compliance review with 25-point checklist for code changes involving payment card data. Produces per-point PASS/FAIL report with remediation.
Generates ai/memory/epic-XXXX-summary.md from a completed epic's artifacts using structured header parsing
Scans Docker images for CVEs and Dockerfile best practices violations. Uses Trivy, Grype, or Snyk Container. Produces SARIF output with scoring.
Stack-aware performance skill: reads QualityConfig.performance SLOs, dispatches to Newman (REST), ghz (gRPC), hyperfine (CLI), Artillery (GraphQL), or custom harness (Socket), compares against governance/baselines/performance-baseline.json, and blocks merge on regression > tolerance.
Runs performance tests to validate latency SLAs, throughput targets, and resource stability under load. Supports baseline, normal, peak, and sustained scenarios.
Generate a Quarkus RESTEasy Reactive @Path resource with DTOs, @RegisterForReflection, ExceptionMapper, and @QuarkusTest unit tests.
Generate a Spring Boot @RestController with matching DTOs, mappers, @ControllerAdvice handler, and unit tests following hexagonal architecture.
Scaffold a Helidon SE or MP service with routing, health checks, config, Dockerfile, and integration tests.
Thin orchestrator (~320 lines — story-0049-0019 refactor) that drives a story end-to-end via 4 delegated phases: Phase 0 (args via x-internal-normalize-args + context via x-internal-load-story-context + resume via x-internal-resume-story), Phase 1 (parallel planning via x-internal-build-story-plan), Phase 2 (task execution loop via x-implement-task per task, then final story PR via x-create-pr), Phase 3 (verify via x-internal-verify-story + report via x-internal-write-story-report + optional worktree cleanup via x-manage-worktrees). New EPIC-0049 flags --target-branch / --auto-merge / --epic-id propagate OO-style to x-implement-task and x-create-pr. Backward compatible: absent flags preserve legacy EPIC-0048 behavior (target=develop, auto-merge=none).
Thin orchestrator (~460 lines — story-0049-0018 refactor) that drives an epic end-to-end via 6 delegated phases: Phase 0 (args via x-internal-normalize-args), Phase 1 (load+plan via x-internal-build-epic-plan), Phase 2 (epic branch via x-internal-ensure-epic-branch), Phase 3 (sequential-by-default story loop via x-implement-story), Phase 4 (integrity gate + report via x-internal-verify-epic-integrity + x-internal-write-report), Phase 5 (final PR epic/XXXX → develop via x-merge-branches + x-create-pr). Defaults flipped by EPIC-0049: sequential execution (opt-in parallel via --parallel), auto-merge of story PRs into epic/XXXX (target changed from develop). Legacy EPIC-0042 behavior preserved under --legacy-flow (auto-detected via execution-state.json flowVersion=1). Zero inline git/gh/jq/mvn calls — orchestrator uses only Read/Glob + Skill.
Implements a feature/story/task using TDD (Red-Green-Refactor) workflow. Schema-aware: v1 (legacy) runs the original Double-Loop TDD flow with story-section task extraction; v2 (task-first, EPIC-0038) reads task-TASK-XXXX-YYYY-NNN.md + plan-task-TASK-XXXX-YYYY-NNN.md, honours declared I/O contracts, respects task-implementation-map dependencies, verifies post-conditions via grep/assert, and produces a single atomic commit per task via x-commit-changes.
Generate Implementation Map from Epic and Stories: dependency matrix, phase computation, critical path, ASCII diagrams, Mermaid graphs, strategic observations. Invoked only by x-create-feature (Phase 4).
Cleans local git state in one pass: fetches origin with prune, removes all non-main worktrees (any path), and deletes all local branches except main/master/develop. Destructive by default with an interactive y/N confirmation gate; supports --dry-run (preview) and --yes (non-interactive).
Detects the resumable state of an in-flight story: reads ai/epics/epic-XXXX/execution-state.json via x-internal-update-status --read-only, identifies the first PENDING or IN_PROGRESS task (the resume point), catalogues DONE tasks with their commitSha, extracts the last committed SHA, and flags staleness when the story file's mtime is newer than any DONE task's completion timestamp. Emits a single-line JSON envelope {resumePoint, tasksCompleted, tasksPending, lastCommitSha, staleWarnings}. Seventh skill in the x-internal-* convention and the fourth under internal/plan/ (after x-internal-load-story-context, x-internal-build-story-plan, x-internal-verify-story). Read-only by construction — never mutates state.
Orchestrates parallel story-planning (Phase 1 carve-out of x-implement-story): invokes x-plan-architecture (Step 1A) and then dispatches 5 sibling Agent subagents in ONE assistant message for implementation plan, test plan, task breakdown, security assessment, and compliance assessment (Steps 1B-1F). Applies the Rule 13 SUBAGENT-GENERAL pattern as the canonical parallel-planning gateway. Scope-aware: SIMPLE skips 1E/1F. Returns a consolidated envelope of artifact paths to the calling orchestrator. Fifth skill in the x-internal-* convention and the second under internal/plan/ (after x-internal-load-story-context).
Changelog generator v2: hybrid format combining narrative Highlights block (derived from 'Entrega de Valor' of merged v2 epics) with Keep-a-Changelog sections (Added/Changed/Fixed/Breaking/Deprecated). Stack-aware via documentation.changelog.format config. Falls back gracefully when EPIC-0070 epics unavailable (D-R10).
Audits compliance of all project rules AND knowledge packs against source code. Launches parallel subagents (one per rule/knowledge-pack) for scanning, then aggregates into a unified report with severity classification and story suggestions.
Build gate verification between parallelism groups. Compiles code, classifies errors, decides retry vs escalate, extracts outputs for next group. Used between each implementation group in Phase 2.
Reconciles execution-state.json (telemetry) against the **Status:** field of Epic / Story markdown artifacts. Default mode (diagnose) is read-only and prints a divergence table. Opt-in --apply rewrites the markdowns atomically via StatusFieldParser and commits via x-commit-changes. Respects Rule 19 (legacy v1 epics skip silently) and Rule 22 (markdown is SoT; state.json is telemetry). Use for manual recovery of legacy epics whose markdown status drifted from execution checkpoints.
Guides incident response with severity-based checklists, communication templates, and postmortem triggers. Interactive guide for SEV1-SEV4 incidents covering classification, response coordination, and action item tracking.
Documentation freshness gate: validates 6 dimensions (readme, api-specs, grpc-proto, adr, skill-docs, system-architecture) against code changes in a PR. Stack-aware — only validates targets declared/auto-detected from ProjectConfig.documentation.targets. Returns exit non-zero on staleness; produces structured report.
Fonte da verdade do modelo RA9 (Rule-Aligned 9-Section) para templates de planejamento Epic/Story/Task. Define as 9 seções fixas, granularidade por nível, micro-template Decision Rationale, e mapeamento rule ↔ seção.
Diagnoses errors, stacktraces, build failures, and unexpected behavior. Systematic approach: reproduce, locate, understand, fix, verify. Use whenever something fails: compilation errors, test failures, runtime exceptions, coverage gaps, or performance issues.
Create Jira Stories from existing local story markdown files. Read all story files in an epic directory, map fields to Jira, create issues with parent epic link, create dependency links between stories, and sync Jira keys back to local files.
Automated profiling: detect language/runtime, select appropriate profiler, execute session, generate flamegraph, identify hotspots, and suggest optimizations referencing the performance-engineering knowledge pack.
Detects and classifies file-collision risks between work units (tasks, stories, or epic phases) before parallel execution. Reads ## File Footprint blocks produced by x-plan-task / x-plan-story, applies the parallelism-heuristics knowledge pack (hard / regen / soft categories + hotspot overrides), and emits a collision matrix + serialization recommendation in Markdown (default) or JSON.
Data modeling specialist review: validates entity design, aggregate boundaries, value objects, repository patterns, domain event persistence, and DDD tactical patterns.
Validates event schemas, producer/consumer patterns, error handling, dead letter topics, and operational readiness for event-driven architectures.
Validates gRPC service definitions, proto3 conventions, implementation patterns, and operational readiness.
Reviews code changes for compliance with selected security frameworks. Verifies sensitive data handling, audit trails, and access control patterns.
Scans Kubernetes manifests, Terraform modules, Helm charts, and Docker Compose files for misconfigurations against CIS benchmarks.
Integrates with SonarQube/SonarCloud for security hotspot tracking, quality gate enforcement, and SARIF output from findings.
Scans Docker images for CVEs and Dockerfile best practices violations. Uses Trivy, Grype, or Snyk Container. Produces SARIF output with scoring.
Scans code and git history for leaked credentials, API keys, tokens, and secrets. Produces SARIF output with scoring and baseline support.
Validates project dependencies against the declared dependency policy (denied CVEs, license whitelist, version constraints, freshness window, scope policy). Produces a structured validation report and exits with BLOCK or WARN based on D-R10/D-R11 enforcement matrix.
Runs automated smoke tests against the REST API using Newman/Postman. Supports local, container-orchestrated, and staging environments.
Stack-aware contract breaking-change detection: openapi-diff (REST), buf breaking (gRPC/proto3), Spring Cloud Contract (Java/Spring), schema registry compat (events). Pact opt-in via quality.contract.pact=true.
Runs integration tests that validate the complete flow from request through all application layers to response, using a real database.
Stack-aware mutation testing skill: detects language/framework, dispatches to PIT (Java/Maven or Gradle), Stryker (JS/TS), mutmut (Python), or go-mutesting (Go), enforces quality.mutation.threshold (default 80%) and runtime-cap-min, and blocks merge on score below threshold.
Regression-shell gate: reads QualityConfig.regression, runs curated scenario scripts against the target (self mode: generator output; service mode: client services), compares against baseline, and blocks merge on unexpected diffs.
Formats source code using the appropriate formatter for {{LANGUAGE}}. First step of the pre-commit chain (format -> lint -> compile -> commit). Supports --check (dry-run) and --changed-only modes.
Generate a Picocli @Command with subcommands, @Option/@Parameters, type converters, exit code constants, and unit tests.
Generate a Quarkus RESTEasy Reactive @Path resource with DTOs, @RegisterForReflection, ExceptionMapper, and @QuarkusTest unit tests.
Detects spec-code drift by comparing story data contracts, endpoints, and Gherkin scenarios against implemented code. Supports standalone mode (full report) and inline mode (compact output for TDD loop integration in x-implement-story Phase 2).
Thin orchestrator (~460 lines — story-0049-0018 refactor) that drives an epic end-to-end via 6 delegated phases: Phase 0 (args via x-internal-normalize-args), Phase 1 (load+plan via x-internal-build-epic-plan), Phase 2 (epic branch via x-internal-ensure-epic-branch), Phase 3 (sequential-by-default story loop via x-implement-story), Phase 4 (integrity gate + report via x-internal-verify-epic-integrity + x-internal-write-report), Phase 5 (final PR epic/XXXX → develop via x-merge-branches + x-create-pr). Defaults flipped by EPIC-0049: sequential execution (opt-in parallel via --parallel), auto-merge of story PRs into epic/XXXX (target changed from develop). Legacy EPIC-0042 behavior preserved under --legacy-flow (auto-detected via execution-state.json flowVersion=1). Zero inline git/gh/jq/mvn calls — orchestrator uses only Read/Glob + Skill.
Thin orchestrator (~320 lines — story-0049-0019 refactor) that drives a story end-to-end via 4 delegated phases: Phase 0 (args via x-internal-normalize-args + context via x-internal-load-story-context + resume via x-internal-resume-story), Phase 1 (parallel planning via x-internal-build-story-plan), Phase 2 (task execution loop via x-implement-task per task, then final story PR via x-create-pr), Phase 3 (verify via x-internal-verify-story + report via x-internal-write-story-report + optional worktree cleanup via x-manage-worktrees). New EPIC-0049 flags --target-branch / --auto-merge / --epic-id propagate OO-style to x-implement-task and x-create-pr. Backward compatible: absent flags preserve legacy EPIC-0048 behavior (target=develop, auto-merge=none).
Analyzes project tech stack and recommends relevant MCP (Model Context Protocol) servers. Auto-detects language, framework, database, cache, and message broker from project config, then matches against a built-in catalog of MCP servers with installation instructions.
Cleans local git state in one pass: fetches origin with prune, removes all non-main worktrees (any path), and deletes all local branches except main/master/develop. Destructive by default with an interactive y/N confirmation gate; supports --dry-run (preview) and --yes (non-interactive).
Creates Conventional Commits with Task ID in scope and pre-commit chain (format -> lint -> compile). Central commit point in the task-centric workflow with TDD tag support.
Manages git worktrees for parallel task and story execution. Operations: create, list, remove, cleanup, detect-context. Follows Rule 14 (Worktree Lifecycle) naming convention under .claude/worktrees/{identifier}/.
Git operations: branch creation, atomic commits (Conventional Commits), push, and PR creation. Use for any git workflow task including branching, committing, pushing, creating PRs, or managing version control.
Generates ai/memory/epic-XXXX-summary.md from a completed epic's artifacts using structured header parsing
Renders _TEMPLATE-*.md templates by substituting {{KEY}} placeholders (simple and nested, dot-path) and resolving {{#each}} loops against a structured JSON data payload, then writes the result atomically to an output path. Supports an --append mode with per-section deduplication keyed by `## ID: <value>` markers. Centralises phase reports, epic execution reports and planning reports so orchestrators (x-implement-epic, x-orchestrate-epic, x-implement-story) stop duplicating `Read template + inline Edit/Write` logic. Second skill in the x-internal-* convention (after x-internal-update-status pilot): internal visibility, non-user-invocable, subdir scoping under internal/ops/.
Builds the canonical ExecutionPlan for an epic (Phase 0/0.5 carve-out of x-implement-epic): loads epic-XXXX.md, IMPLEMENTATION-MAP.md, and every story-*.md; constructs the inter-story dependency DAG; runs Kahn's algorithm with cycle detection; optionally computes a file-overlap matrix (mode=parallel) and the critical path; then renders ai/epics/epic-XXXX/epic-execution-plan.md via x-internal-write-report using the _TEMPLATE-EPIC-EXECUTION-PLAN.md template. Emits a stable JSON envelope on stdout for orchestrator consumption. Sixth skill in the x-internal-* convention and the third under internal/plan/ (after x-internal-load-story-context and x-internal-build-story-plan).
Orchestrates parallel story-planning (Phase 1 carve-out of x-implement-story): invokes x-plan-architecture (Step 1A) and then dispatches 5 sibling Agent subagents in ONE assistant message for implementation plan, test plan, task breakdown, security assessment, and compliance assessment (Steps 1B-1F). Applies the Rule 13 SUBAGENT-GENERAL pattern as the canonical parallel-planning gateway. Scope-aware: SIMPLE skips 1E/1F. Returns a consolidated envelope of artifact paths to the calling orchestrator. Fifth skill in the x-internal-* convention and the second under internal/plan/ (after x-internal-load-story-context).
Generate Story files from Epic: data contracts, Gherkin, Mermaid diagrams, dependency declarations, sub-tasks, quality validation, optional Jira integration. Invoked only by x-create-feature (Phase 3).
Documentation automation v2: stack-aware generation consuming documentation.targets from ProjectConfig. Detects documentation type needed (API, README, ADR, changelog, system-architecture) from code changes and stack config, delegates to specialized skills or generates inline. Invokes x-update-system-architecture on architectural change detection.
Validates phase transitions for orchestrators under Rule 25. Four modes: --mode pre (assert predecessor phases completed before entering phase N), --mode post (assert all child tasks of phase N are completed AND all expected artifacts exist on disk), --mode wave (post-Batch-B verification of parallel wave completeness: N child TaskUpdate completed + N artifacts exist), --mode final (terminal gate composing with x-internal-verify-epic-integrity). Reads execution-state.json.taskTracking.phaseGateResults and TaskList task state; writes back the gate result. Emits a single-line JSON envelope {passed, mode, skill, phase, expectedTasks, completedTasks, missingTasks, expectedArtifacts, missingArtifacts, wallclockMs, timestamp}. Exit 0 on passed, 12 on failure, 13 on malformed args, 14 on task-resolution timeout. First skill in the x-internal-* convention authored by EPIC-0055; eighth overall (after status-update, report-write, args-normalize, story-load-context, story-build-plan, story-verify, story-resume, epic-build-plan, epic-integrity-gate, epic-branch-ensure, story-report) and the eighth under internal/plan/.
Executes the epic-level verification gate end-to-end (Phase 1.7 / Phase 4 carve-out of x-implement-epic) on the epic/XXXX branch HEAD: checks out the branch, runs mvn clean test + jacoco:report, parses filtered coverage against epic-level thresholds (default line >=95, branch >=90), runs a declarative DoD checklist (presence of tests, tasks DONE, CHANGELOG entry, ADR references), and emits a single-line JSON envelope {passed, failures, coverageDelta, dodChecklist}. Seventh skill in the x-internal-* convention and the fourth under internal/plan/ (after x-internal-load-story-context, x-internal-build-story-plan, and x-internal-verify-story). Isolates ~180 inline lines of integrity-gate logic from x-implement-epic.
Generate Implementation Map from Epic and Stories: dependency matrix, phase computation, critical path, ASCII diagrams, Mermaid graphs, strategic observations. Invoked only by x-create-feature (Phase 4).
Detects the resumable state of an in-flight story: reads ai/epics/epic-XXXX/execution-state.json via x-internal-update-status --read-only, identifies the first PENDING or IN_PROGRESS task (the resume point), catalogues DONE tasks with their commitSha, extracts the last committed SHA, and flags staleness when the story file's mtime is newer than any DONE task's completion timestamp. Emits a single-line JSON envelope {resumePoint, tasksCompleted, tasksPending, lastCommitSha, staleWarnings}. Seventh skill in the x-internal-* convention and the fourth under internal/plan/ (after x-internal-load-story-context, x-internal-build-story-plan, x-internal-verify-story). Read-only by construction — never mutates state.
Renders _TEMPLATE-*.md templates by substituting {{KEY}} placeholders (simple and nested, dot-path) and resolving {{#each}} loops against a structured JSON data payload, then writes the result atomically to an output path. Supports an --append mode with per-section deduplication keyed by `## ID: <value>` markers. Centralises phase reports, epic execution reports and planning reports so orchestrators (x-epic-implement, x-epic-orchestrate, x-story-implement) stop duplicating `Read template + inline Edit/Write` logic. Second skill in the x-internal-* convention (after x-internal-status-update pilot): internal visibility, non-user-invocable, subdir scoping under internal/ops/.
Renders PR body Markdown from existing disk artifacts using one of two templates. --kind=implementation: reads review/verify/telemetry artifacts for a story and emits a structured implementation PR body. --kind=backlog: reads epic/map artifacts and emits a scaffolding PR body. Fail-open: absent artifacts produce human-readable placeholders, never audit-sentinel strings. Internal — invoked only by x-pr-create and x-feature-create.
Create Jira Stories from existing local story markdown files. Read all story files in an epic directory, map fields to Jira, create issues with parent epic link, create dependency links between stories, and sync Jira keys back to local files.
Audits compliance of all project rules AND knowledge packs against source code. Launches parallel subagents (one per rule/knowledge-pack) for scanning, then aggregates into a unified report with severity classification and story suggestions.
Decomposes an implementation plan into tasks. Primary mode: derives tasks from test scenarios (x-plan-tests output) using TDD structure (RED/GREEN/REFACTOR). Fallback mode: uses Layer Task Catalog (G1-G7) when no test plan exists.
Build gate verification between parallelism groups. Compiles code, classifies errors, decides retry vs escalate, extracts outputs for next group. Used between each implementation group in Phase 2.
Detect cross-epic P95 regressions (>= threshold %) and rank top-10 slowest skills from the global telemetry index. Single-responsibility partner of /x-analyze-telemetry focused on trend detection, not point-in-time reporting. Use to answer 'is skill X getting slower over the last N epics?' with evidence.
Analyze telemetry NDJSON for one or more epics and produce a Markdown report with skill/phase/tool aggregates, Mermaid Gantt timeline, and optional JSON/CSV exports. Use to answer 'which phase is the bottleneck?' and 'is skill X getting slower?' questions for operator visibility.
Guides incident response with severity-based checklists, communication templates, and postmortem triggers. Interactive guide for SEV1-SEV4 incidents covering classification, response coordination, and action item tracking.
Documentation freshness gate: validates 6 dimensions (readme, api-specs, grpc-proto, adr, skill-docs, system-architecture) against code changes in a PR. Stack-aware — only validates targets declared/auto-detected from ProjectConfig.documentation.targets. Returns exit non-zero on staleness; produces structured report.
Assists migration of a v1 epic document to the v2 value-driven template (EPIC-0070). Parses v1 technical blocks (Packages, Contratos, SOLID, Observabilidade), classifies each block with a safe default heuristic, and optionally asks the operator for confirmation per block (--interactive). Side-effects: writes epic.md in v2 format atomically, creates ADRs for 'virar ADR' decisions, updates system.md via x-update-system-architecture. Supports --dry-run and recovery from interrupted sessions.
Multi-agent story planning: launches 5 specialized agents (Architect, QA, Security, Tech Lead, Product Owner) in parallel to produce a consolidated task breakdown, individual task plans, planning report, and DoR validation. Schema-aware: v1 (legacy) runs the original 6-phase flow; v2 (task-first, EPIC-0038) adds Phases 4a-4c that emit task-TASK-NNN.md + plan-task-TASK-NNN.md per task and a task-implementation-map-STORY-*.md, wiring every task through x-plan-task in parallel.
Transform free-form prose or a text file into a structured RA9 spec (5 mandatory sections) and open a PR on docs/feature-<slug> targeting develop for human review before invoking x-feature-create.
Generate threat models using STRIDE analysis: identify components, map data flows, analyze threats per category, classify severity, suggest mitigations, and produce threat model document.
Generates a detailed per-task implementation plan (plan-task-TASK-XXXX-YYYY-NNN.md) with TDD cycles in TPP order, file impact analysis by architecture layer, security checklist by task type, and exit criteria. Two invocation modes: task-file-first (--task-file) consumes a standalone task-TASK-XXXX-YYYY-NNN.md contract (EPIC-0038); story-scoped (STORY-ID --task TASK-ID) reads the task from story Section 8 (legacy). Invocable standalone OR via x-plan-story (future).
Incrementally updates the service architecture document with changes from architecture plans. Adds new components, integrations, flows, and ADR references without rewriting existing content. Use after implementation to keep architecture documentation current.
Transform free-form prose or a text file into a structured RA9 spec (5 mandatory sections) and open a PR on docs/feature-<slug> targeting develop for human review before invoking x-create-feature.
Transform free-form prose or a text file into a structured RA9 spec (5 mandatory sections) and open a PR on docs/feature-<slug> targeting develop for human review before invoking x-feature-create.
Multi-persona 4-phase story refinement dispatcher. Phase A: 5-7 parallel specialist agents analyse the story for gaps. Phase B: single consolidated question batch to the operator. Phase C: 5-7 parallel specialists refine with answers. Phase D: Architect (opus) consolidates a Refinement Verdict and dual-writes to execution-state.json + story markdown.
Multi-persona 4-phase strategic epic refinement dispatcher. Phase A: 5-6 parallel specialist agents analyse the epic for strategic gaps. Phase B: single consolidated question batch to the operator. Phase C: 5-6 parallel specialists refine with answers. Phase D: Architect (opus) consolidates a Refinement Verdict with scope=epic and dual-writes to execution-state.json + epic markdown.
Transform free-form prose or a text file into a structured RA9 spec (5 mandatory sections) and open a PR on docs/feature-<slug> targeting develop for human review before invoking x-create-feature.
Runs automated smoke tests against the TCP socket server using a standalone Java client with message framing and protocol validation.
Multi-persona 4-phase story refinement dispatcher. Phase A: 5-7 parallel specialist agents analyse the story for gaps. Phase B: single consolidated question batch to the operator. Phase C: 5-7 parallel specialists refine with answers. Phase D: Architect (opus) consolidates a Refinement Verdict and dual-writes to execution-state.json + story markdown.
Generates a comprehensive architecture plan with component diagrams, sequence diagrams, deployment topology, mini-ADRs, NFRs, and resilience/observability strategies. Use before implementation to document design decisions.
Atomic read-modify-write of execution-state.json with flock-based concurrency, schema validation, and idempotency detection. Substitutes inline Edit-based mutations in orchestrator skills (x-implement-epic, x-implement-story, x-fix-epic-pr) that previously suffered race conditions in --parallel mode. PILOT skill for the x-internal-* convention: internal visibility, non-user-invocable, subdir scoping under internal/ops/.
Validates phase transitions for orchestrators under Rule 25. Four modes: --mode pre (assert predecessor phases completed before entering phase N), --mode post (assert all child tasks of phase N are completed AND all expected artifacts exist on disk), --mode wave (post-Batch-B verification of parallel wave completeness: N child TaskUpdate completed + N artifacts exist), --mode final (terminal gate composing with x-internal-verify-epic-integrity). Reads execution-state.json.taskTracking.phaseGateResults and TaskList task state; writes back the gate result. Emits a single-line JSON envelope {passed, mode, skill, phase, expectedTasks, completedTasks, missingTasks, expectedArtifacts, missingArtifacts, wallclockMs, timestamp}. Exit 0 on passed, 12 on failure, 13 on malformed args, 14 on task-resolution timeout. First skill in the x-internal-* convention authored by EPIC-0055; eighth overall (after status-update, report-write, args-normalize, story-load-context, story-build-plan, story-verify, story-resume, epic-build-plan, epic-integrity-gate, epic-branch-ensure, story-report) and the eighth under internal/plan/.
Analyzes source code with the appropriate linter for {{LANGUAGE}}. Second step in the pre-commit chain (RULE-007: format -> lint -> compile -> commit). Supports --fix, --changed-only, and --strict modes.
Guides incident response with severity-based checklists, communication templates, and postmortem triggers. Interactive guide for SEV1-SEV4 incidents covering classification, response coordination, and action item tracking.
Fonte da verdade do modelo RA9 (Rule-Aligned 9-Section) para templates de planejamento Epic/Story/Task. Define as 9 seções fixas, granularidade por nível, micro-template Decision Rationale, e mapeamento rule ↔ seção.
Generates a detailed per-task implementation plan (plan-task-TASK-XXXX-YYYY-NNN.md) with TDD cycles in TPP order, file impact analysis by architecture layer, security checklist by task type, and exit criteria. Two invocation modes: task-file-first (--task-file) consumes a standalone task-TASK-XXXX-YYYY-NNN.md contract (EPIC-0038); story-scoped (STORY-ID --task TASK-ID) reads the task from story Section 8 (legacy). Invocable standalone OR via x-story-plan (future).
QA specialist review: validates test coverage, TDD compliance, test naming, fixtures, parametrized tests, and acceptance criteria coverage.
Aggregates results from all security scanning skills into a unified posture view with score 0-100, trend tracking, OWASP risk heatmap, per-dimension breakdown, and remediation priority queue. Never executes scans — reads existing results only (RULE-011).
Generate an Implementation Map from an Epic and its Stories with dependency matrix, phase computation, critical path analysis, ASCII phase diagrams, Mermaid dependency graphs, phase summary tables, and strategic observations.
Orchestrates complete release flow using Git Flow release branches with approval gate, PR-flow (gh CLI) and deep validation: version bump (auto-detect or explicit), release branch creation from develop, deep validation (coverage, golden files, version consistency), version file updates, changelog generation, release commit, release PR via gh (optionally reviewed by x-review-pr), human approval gate with persistent state file, tag on main after merged PR, back-merge PR to develop with conflict detection, and cleanup. Supports hotfix releases from main, dry-run mode, resume via --continue-after-merge, in-session pause via --interactive, GPG-signed tags, skip-review opt-out, and custom state file path.
Multi-phase penetration test orchestrator: reconnaissance, vulnerability scanning, exploitation validation, and consolidated reporting with environment-based restrictions.
Detects and classifies file-collision risks between work units (tasks, stories, or epic phases) before parallel execution. Reads ## File Footprint blocks produced by x-task-plan / x-story-plan, applies the parallelism-heuristics knowledge pack (hard / regen / soft categories + hotspot overrides), and emits a collision matrix + serialization recommendation in Markdown (default) or JSON.
Scaffolds a Micronaut service with @Controller, DI, health, Dockerfile, and tests.
Scaffolds a Helidon SE/MP service with routing, health, config, Dockerfile, and tests.
Scaffolds a Helidon SE/MP service with routing, health, config, Dockerfile, and tests.
Thin orchestrator (~320 lines — story-0049-0019 refactor) that drives a story end-to-end via 4 delegated phases: Phase 0 (args via x-internal-args-normalize + context via x-internal-story-load-context + resume via x-internal-story-resume), Phase 1 (parallel planning via x-internal-story-build-plan), Phase 2 (task execution loop via x-task-implement per task, then final story PR via x-pr-create), Phase 3 (verify via x-internal-story-verify + report via x-internal-story-report + optional worktree cleanup via x-git-worktree). New EPIC-0049 flags --target-branch / --auto-merge / --epic-id propagate OO-style to x-task-implement and x-pr-create. Backward compatible: absent flags preserve legacy EPIC-0048 behavior (target=develop, auto-merge=none).
Thin orchestrator (~460 lines — story-0049-0018 refactor) that drives an epic end-to-end via 6 delegated phases: Phase 0 (args via x-internal-args-normalize), Phase 1 (load+plan via x-internal-epic-build-plan), Phase 2 (epic branch via x-internal-epic-branch-ensure), Phase 3 (sequential-by-default story loop via x-story-implement), Phase 4 (integrity gate + report via x-internal-epic-integrity-gate + x-internal-report-write), Phase 5 (final PR epic/XXXX → develop via x-git-merge + x-pr-create). Defaults flipped by EPIC-0049: sequential execution (opt-in parallel via --parallel), auto-merge of story PRs into epic/XXXX (target changed from develop). Legacy EPIC-0042 behavior preserved under --legacy-flow (auto-detected via execution-state.json flowVersion=1). Zero inline git/gh/jq/mvn calls — orchestrator uses only Read/Glob + Skill.
Generates a comprehensive architecture plan with component diagrams, sequence diagrams, deployment topology, mini-ADRs, NFRs, and resilience/observability strategies. Use before implementation to document design decisions.
Task-level PR creation with formatted title, automatic labels, structured body, and target branch logic. Creates standardized PRs for individual tasks with Task ID traceability.
Bumps the application version following Semantic Versioning (patch/minor/major) and creates a conventional commit.
Generates a Double-Loop TDD test plan with TPP-ordered scenarios before implementation. Delegates KP reading to a context-gathering subagent, then produces structured Acceptance Tests (outer loop) and Unit Tests in Transformation Priority Premise order (inner loop).
Validate and configure local development environment: detect stack, check prerequisites, verify versions, validate IDE config, test database connectivity, run initial build, and report status with fix suggestions.
Reconciles execution-state.json (telemetry) against the **Status:** field of Epic / Story markdown artifacts. Default mode (diagnose) is read-only and prints a divergence table. Opt-in --apply rewrites the markdowns atomically via StatusFieldParser and commits via x-commit-changes. Respects Rule 19 (legacy v1 epics skip silently) and Rule 22 (markdown is SoT; state.json is telemetry). Use for manual recovery of legacy epics whose markdown status drifted from execution checkpoints.
Generates the final consolidated story-completion report by reading ai/epics/epic-XXXX/execution-state.json, collecting per-task status and commitSha, PR metadata (prNumber, prState), coverage delta, and review findings, then rendering the output via x-internal-write-report with _TEMPLATE-STORY-COMPLETION-REPORT.md to the caller-specified --output path. Eighth skill in the x-internal-* convention and the fifth under internal/plan/ (after x-internal-load-story-context, x-internal-build-story-plan, x-internal-verify-story, and x-internal-resume-story). Read-only against state; writes only to --output via x-internal-write-report.
Dockerfile patterns: multi-stage builds, security hardening, .dockerignore templates, layer optimization, health checks, and OCI labels per language.
Incrementally updates the service architecture document with changes from architecture plans. Adds new components, integrations, flows, and ADR references without rewriting existing content. Use after implementation to keep architecture documentation current.
Reviews code changes for compliance with selected security frameworks. Verifies sensitive data handling, audit trails, and access control patterns.
Executes the story-level verification gate (Phase 3 carve-out of x-implement-story): identifies files touched by the story via the task breakdown, runs the test suite scoped to those files, parses filtered coverage against the story-specific thresholds (default line >=95, branch >=90), performs cross-file consistency checks (constructor patterns, return-type uniformity per role), optionally runs the smoke suite, and validates every Section 7 Gherkin scenario has a matching acceptance test. Emits a single-line JSON envelope {passed, coverageDelta, failures, acCheckResults}. Sixth skill in the x-internal-* convention and the third under internal/plan/ (after x-internal-load-story-context and x-internal-build-story-plan).
Tech Lead review with dynamic architecture-driven dimensions; reads all specialist and developer reports to issue GO/NO-GO.
Generates a Double-Loop TDD test plan with TPP-ordered acceptance and unit scenarios.
Builds a deterministic hybrid tech profile from config, file/diff auto-detection, and caller context.
Migrates artifact frontmatter from v2 to v3.0 (requires-capabilities) using path heuristics, keyword scan, and AI fallback
Diagnoses errors, stacktraces, build failures, and unexpected behavior. Systematic approach: reproduce, locate, understand, fix, verify. Use whenever something fails: compilation errors, test failures, runtime exceptions, coverage gaps, or performance issues.
Formats source code; first step of the pre-commit chain (format -> lint -> compile).
Renders _TEMPLATE-*.md templates by substituting {{KEY}} placeholders (simple and nested, dot-path) and resolving {{#each}} loops against a structured JSON data payload, then writes the result atomically to an output path. Supports an --append mode with per-section deduplication keyed by `## ID: <value>` markers. Centralises phase reports, epic execution reports and planning reports so orchestrators (x-epic-implement, x-epic-orchestrate, x-story-implement) stop duplicating `Read template + inline Edit/Write` logic. Second skill in the x-internal-* convention (after x-internal-status-update pilot): internal visibility, non-user-invocable, subdir scoping under internal/ops/.
Documentation automation v2: stack-aware generation from documentation.targets.
Task-level PR creation with formatted title, labels, structured body, and target branch.
Transform an ideation file into a Product artifact with RNF roots and C1 capability stub.
Knowledge pack for spec-driven planning concepts: Planning Depth Tier, Gray-Area Discovery, Knowledge Verification Chain, Requirement Traceability, and the Deferred-Ideas ledger. Subordinate to Rule 27 (zero-bypass).
Generates a Spring Boot @RestController with DTOs, mappers, advice, and unit tests.
Scans Kubernetes/Terraform/Helm/Compose files for misconfigurations vs CIS benchmarks.
Supply-chain audit: maintainer risk, typosquatting, EPSS, SLSA; SARIF + report.
Batch-commits planning artifacts under plans/** without the code pre-commit chain.
Evaluates runtime protection (rate limits, WAF, CSP) with SARIF + ASVS scoring.
Runs Newman/Postman smoke tests against the REST API across local/container/staging.
Runs integration tests covering the full request flow with a real database.
Discovers all PRs from an epic, classifies comments, applies fixes, opens correction PR.
Runs automated smoke tests against the REST API using Newman/Postman. Supports local, container-orchestrated, and staging environments.
Conditional performance-tests gate: SLA budget validation per quality.performance.
Scaffolds a Helidon SE/MP service with routing, health, config, Dockerfile, and tests.
Scaffolds a Micronaut service with @Controller, DI, health, Dockerfile, and tests.
Generates a Quarkus RESTEasy Reactive @Path resource with DTOs, mapper, and tests.
Executes Red-Green-Refactor TDD cycles for a task; delegates commits to x-commit-changes.
Generates or updates CI/CD pipelines per project stack with actionlint validation.
Regression-shell gate: runs scenario scripts vs baseline; blocks merge on diffs.
Scaffolds a Helidon SE/MP service with routing, health, config, Dockerfile, and tests.
Decompose a Product into Capabilities with explicit RNF inheritance and no-relax markers.
Full codebase review against project standards via parallel specialist subagents.
Conditional mutation-testing gate (PIT/Stryker) with kill-rate threshold.
Builds the canonical ExecutionPlan for an epic (Phase 0/0.5 carve-out of x-epic-implement): loads epic-XXXX.md, IMPLEMENTATION-MAP.md, and every story-*.md; constructs the inter-story dependency DAG; runs Kahn's algorithm with cycle detection; optionally computes a file-overlap matrix (mode=parallel) and the critical path; then renders plans/epic-XXXX/epic-execution-plan.md via x-internal-report-write using the _TEMPLATE-EPIC-EXECUTION-PLAN.md template. Emits a stable JSON envelope on stdout for orchestrator consumption. Sixth skill in the x-internal-* convention and the third under internal/plan/ (after x-internal-story-load-context and x-internal-story-build-plan).
Creates a focused Epic artifact from a Feature markdown, preserving lineage and RNFs.
Executes the story-level verification gate (Phase 3 carve-out of x-implement-story): identifies files touched by the story via the task breakdown, runs the test suite scoped to those files, parses filtered coverage against the story-specific thresholds (default line >=95, branch >=90), performs cross-file consistency checks (constructor patterns, return-type uniformity per role), optionally runs the smoke suite, and validates every Section 7 Gherkin scenario has a matching acceptance test. Emits a single-line JSON envelope {passed, coverageDelta, failures, acCheckResults}. Sixth skill in the x-internal-* convention and the third under internal/plan/ (after x-internal-load-story-context and x-internal-build-story-plan).
Evaluates application hardening posture against CIS and OWASP benchmarks: HTTP security headers, TLS configuration, CORS policy, cookie security, error handling, input limits, and information disclosure. Produces SARIF output with weighted scoring.
Closes a feature: validates all stories are MERGED into feature/XXXX, syncs develop, opens the final PR to develop, and generates the memory summary.
Scans code and git history for leaked credentials, API keys, and tokens; SARIF output.
Disaster recovery patterns: DR strategies, RPO/RTO, failover automation, DR testing, multi-region patterns, and recovery procedures
Documentation freshness gate: validates 6 dimensions (readme, api, adr, etc.) per PR.
Discovers all PRs from an epic via execution-state.json, fetches and classifies review comments in batch, generates a consolidated findings report, applies fixes, and creates a single correction PR. Supports dry-run, explicit PR list fallback, and idempotent re-execution.
Creates Jira Stories from local story markdowns; links parent Epic and dependencies.
Changelog generator v2: hybrid format combining narrative Highlights block (derived from 'Entrega de Valor' of merged v2 epics) with Keep-a-Changelog sections (Added/Changed/Fixed/Breaking/Deprecated). Stack-aware via documentation.changelog.format config. Falls back gracefully when EPIC-0070 epics unavailable (D-R10).
Database specialist review: dynamically loads versioned patterns for detected DB engine and version.
Implements a feature/story/task using TDD (Red-Green-Refactor) workflow. Schema-aware: v1 (legacy) runs the original Double-Loop TDD flow with story-section task extraction; v2 (task-first, EPIC-0038) reads task-TASK-XXXX-YYYY-NNN.md + plan-task-TASK-XXXX-YYYY-NNN.md, honours declared I/O contracts, respects task-implementation-map dependencies, verifies post-conditions via grep/assert, and produces a single atomic commit per task via x-git-commit.
Quarkus patterns: CDI, @ConfigMapping, Panache Repository, RESTEasy Reactive, native build constraints, and @RegisterForReflection.
Formats source code using the appropriate formatter for {{LANGUAGE}}. First step of the pre-commit chain (format -> lint -> compile -> commit). Supports --check (dry-run) and --changed-only modes.
Creates Conventional Commits with Task ID and pre-commit chain (format/lint/compile).
Generates a Picocli @Command with subcommands, options, converters, and unit tests.
Validates gRPC service definitions, proto3 conventions, implementation patterns, and operational readiness.
Detects spec-code drift by comparing story data contracts, endpoints, and Gherkin scenarios against implemented code. Supports standalone mode (full report) and inline mode (compact output for TDD loop integration in x-implement-story Phase 2).
Generates a detailed per-task implementation plan (plan-task-TASK-XXXX-YYYY-NNN.md) with TDD cycles in TPP order, file impact analysis by architecture layer, security checklist by task type, and exit criteria. Two invocation modes: task-file-first (--task-file) consumes a standalone task-TASK-XXXX-YYYY-NNN.md contract (EPIC-0038); story-scoped (STORY-ID --task TASK-ID) reads the task from story Section 8 (legacy). Invocable standalone OR via x-story-plan (future).
Validates GraphQL schema design, resolver implementation, security patterns, and observability for compliance with best practices.
Adds or reviews distributed tracing, metrics, and structured logging using OpenTelemetry SDK with OTLP export.
DAST gate: ZAP + Nuclei (passive/active); SARIF + Markdown; blocks on HIGH/CRITICAL.
Tech Lead holistic review with {review_max_score}-point checklist covering Clean Code, SOLID, architecture, framework conventions, tests, TDD process, security, and cross-file consistency. Produces GO/NO-GO decision. Use for final review before merge.
PCI-DSS compliance review with 15-point checklist for payment-card-data changes.
Executes systematic Red-Green-Refactor TDD cycles for a task. Reads the task plan generated by x-task-plan, runs each cycle in TPP order, validates RED/GREEN/REFACTOR phases, delegates atomic commits to x-git-commit with TDD tags, and supports resume and dry-run.
Observability principles: distributed tracing (span trees, mandatory attributes), metrics naming conventions, structured logging with mandatory fields, health checks (liveness/readiness/startup), correlation IDs, and OpenTelemetry integration.
Decomposes an implementation plan into tasks. Primary mode: derives tasks from test scenarios (x-plan-tests output) using TDD structure (RED/GREEN/REFACTOR). Fallback mode: uses Layer Task Catalog (G1-G7) when no test plan exists.
Architecture pattern references: microservice, resilience, data, integration, and architectural patterns (saga, outbox, circuit breaker, CQRS, event sourcing, and more).
Multi-agent story planning: 7 specialized agents produce task breakdown and plans.
Canonical catalog of file-collision heuristics for parallel task, story, and epic execution. Defines the File Footprint block format, the three Conflict Categories (hard / regen / soft), a hotspot list of high-contention paths, and the degrade-with-warning policy. Consumed by x-task-plan, x-story-plan, x-epic-map, x-parallel-eval, and x-dev-*-implement.
Validate and configure local development environment: detect stack, check prerequisites, verify versions, validate IDE config, test database connectivity, run initial build, and report status with fix suggestions.
OWASP ASVS 4.0.3 verification standard with L1/L2/L3 levels, cross-reference tables, and verification items for all 14 chapters (V1-V14).
Parses an argv string against a declarative JSON schema with typed flags (boolean, string, integer, enum), defaults, mutually-exclusive groups, and deprecation warnings, then emits a normalized `{parsed, warnings, errors}` envelope on stdout. Replaces ~150 lines of inline argv parsing inlined inside `x-epic-implement`, `x-story-implement`, and `x-epic-orchestrate`, giving every orchestrator identical flag-validation syntax and error messages. Third skill in the x-internal-* convention (after x-internal-status-update pilot and x-internal-report-write): internal visibility, non-user-invocable, subdir scoping under internal/ops/.
Stack-aware performance skill: reads QualityConfig.performance SLOs, dispatches to Newman (REST), ghz (gRPC), hyperfine (CLI), Artillery (GraphQL), or custom harness (Socket), compares against governance/baselines/performance-baseline.json, and blocks merge on regression > tolerance.
Database specialist review: validates schema design, migration safety, query optimization, connection management, transaction boundaries, and data integrity patterns.
Detect cross-epic P95 regressions (>= threshold %) and rank top-10 slowest skills from the global telemetry index. Single-responsibility partner of /x-analyze-telemetry focused on trend detection, not point-in-time reporting. Use to answer 'is skill X getting slower over the last N epics?' with evidence.
Validates GraphQL schema design, resolver implementation, security patterns, and observability for compliance with best practices.
Reviews event schemas, producer/consumer patterns, error handling, and DLQ readiness.
Validates phase transitions for orchestrators (Rule 25): pre/post/wave/final modes.
Spring Boot patterns: DI, @ConfigurationProperties, Spring Data JPA, @RestController, @ControllerAdvice, Spring AOT, and health checks.
Analyze telemetry NDJSON for one or more epics and produce a Markdown report with skill/phase/tool aggregates, Mermaid Gantt timeline, and optional JSON/CSV exports. Use to answer 'which phase is the bottleneck?' and 'is skill X getting slower?' questions for operator visibility.
DevOps specialist review: validates Dockerfile, container security, CI/CD pipeline, resource limits, health probes, graceful shutdown, and deployment configuration.
Decomposes an implementation plan into tasks. Primary mode: derives tasks from test scenarios (x-test-plan output) using TDD structure (RED/GREEN/REFACTOR). Fallback mode: uses Layer Task Catalog (G1-G7) when no test plan exists.
Kubernetes deployment patterns: workload types, pod specifications, resource sizing, probes, autoscaling, network policies, and security contexts.
Evaluates application hardening posture against CIS and OWASP benchmarks: HTTP security headers, TLS configuration, CORS policy, cookie security, error handling, input limits, and information disclosure. Produces SARIF output with weighted scoring.
DDD Strategic Design: bounded context identification, context map with 6 integration patterns, Anti-Corruption Layer template, and /x-ddd-context-map skill.
Integrates with SonarQube/SonarCloud for security hotspot tracking, quality gate enforcement, and SARIF output from findings.
Promote interactive ideation output from x-create-feature to .aikittools/ideations/.
Multi-persona 4-phase strategic feature refinement: parallel specialists + verdict. On approval, auto-creates the feature container.
CQRS/Event Sourcing patterns: write/read model separation, command bus, event store, aggregate with event sourcing, projections, snapshot policy, and dead letter handling.
API design principles: {{LANGUAGE}}-specific patterns for REST/gRPC/GraphQL. URL structure, status codes, RFC 7807 errors, pagination, content negotiation, validation, request/response shaping, versioning strategies, and protocol conventions.
Data modeling specialist review: validates entity design, aggregate boundaries, value objects, repository patterns, domain event persistence, and DDD tactical patterns.
Data management lifecycle patterns: zero-downtime migrations, expand/contract pattern, schema versioning, data governance, backup/restore strategies, partitioning, CDC, and data quality validation for {{DATABASE_TYPE}} with {{MIGRATION_TOOL}}.
Orchestrates Git Flow release: version bump, validation, PR, tag, back-merge.
Database specialist review: validates schema design, migration safety, query optimization, connection management, transaction boundaries, and data integrity patterns.
Executes the epic-level verification gate end-to-end (Phase 1.7 / Phase 4 carve-out of x-epic-implement) on the epic/XXXX branch HEAD: checks out the branch, runs mvn clean test + jacoco:report, parses filtered coverage against epic-level thresholds (default line >=95, branch >=90), runs a declarative DoD checklist (presence of tests, tasks DONE, CHANGELOG entry, ADR references), and emits a single-line JSON envelope {passed, failures, coverageDelta, dodChecklist}. Seventh skill in the x-internal-* convention and the fourth under internal/plan/ (after x-internal-story-load-context, x-internal-story-build-plan, and x-internal-story-verify). Isolates ~180 inline lines of integrity-gate logic from x-epic-implement.
Generate Story files from Epic: data contracts, Gherkin, Mermaid diagrams, dependency declarations, sub-tasks, quality validation, optional Jira integration. Invoked only by x-create-feature (Phase 3).
Runs tests with coverage reporting and threshold validation. Use whenever writing, running, or analyzing tests. Triggers on: test, coverage, TDD, unit test, integration test, test failure, coverage gap, or Definition of Done validation.
Full codebase review against project standards via parallel specialist subagents.
Scans Kubernetes manifests, Terraform modules, Helm charts, and Docker Compose files for misconfigurations against CIS benchmarks.
Helm chart patterns: chart structure, values templates, multi-environment configuration, dependencies, testing, GitOps integration, and Helmfile.
Static Application Security Testing -- scans source code for security vulnerabilities without executing the application. Produces SARIF output with OWASP mapping.
SRE practices: error budgets, toil reduction, on-call practices, capacity planning, incident management process, and change management. Covers SLO-based release gates, burn rate calculation, automation prioritization, rotation patterns, load testing methodology, blameless postmortems, and canary analysis.
Stack-aware contract breaking-change detection: openapi-diff (REST), buf breaking (gRPC/proto3), Spring Cloud Contract (Java/Spring), schema registry compat (events). Pact opt-in via quality.contract.pact=true.
Generates a Picocli @Command with subcommands, options, converters, and unit tests.
Multi-agent story planning: launches 5 specialized agents (Architect, QA, Security, Tech Lead, Product Owner) in parallel to produce a consolidated task breakdown, individual task plans, planning report, and DoR validation. Schema-aware: v1 (legacy) runs the original 6-phase flow; v2 (task-first, EPIC-0038) adds Phases 4a-4c that emit task-TASK-NNN.md + plan-task-TASK-NNN.md per task and a task-implementation-map-STORY-*.md, wiring every task through x-plan-task in parallel.
Documentation automation v2: stack-aware generation consuming documentation.targets from ProjectConfig. Detects documentation type needed (API, README, ADR, changelog, system-architecture) from code changes and stack config, delegates to specialized skills or generates inline. Invokes x-update-system-architecture on architectural change detection.
Merges a source branch into a target branch locally with configurable strategy (merge/squash/rebase), automatic conflict detection + rollback, and idempotent no-op when target already contains source HEAD. Centralizes the ~120 lines of inline Bash previously in x-implement-epic Phase 1.4e auto-rebase.
Validate and configure local development environment: detect stack, check prerequisites, verify versions, validate IDE config, test database connectivity, run initial build, and report status with fix suggestions.
Generates ADRs from architecture-plan mini-ADRs with sequential numbering and index update.
Parses an argv string against a declarative JSON schema with typed flags (boolean, string, integer, enum), defaults, mutually-exclusive groups, and deprecation warnings, then emits a normalized `{parsed, warnings, errors}` envelope on stdout. Replaces ~150 lines of inline argv parsing inlined inside `x-epic-implement`, `x-story-implement`, and `x-epic-orchestrate`, giving every orchestrator identical flag-validation syntax and error messages. Third skill in the x-internal-* convention (after x-internal-status-update pilot and x-internal-report-write): internal visibility, non-user-invocable, subdir scoping under internal/ops/.
Generate a Picocli @Command with subcommands, @Option/@Parameters, type converters, exit code constants, and unit tests.
Reviews API gateway configuration for routing rules, authentication, rate limiting, CORS, security headers, TLS, and observability integration.
Scans code and git history for leaked credentials, API keys, tokens, and secrets. Produces SARIF output with scoring and baseline support.
Multi-persona 4-phase strategic epic refinement dispatcher. Phase A: 5-6 parallel specialist agents analyse the epic for strategic gaps. Phase B: single consolidated question batch to the operator. Phase C: 5-6 parallel specialists refine with answers. Phase D: Architect (opus) consolidates a Refinement Verdict with scope=epic and dual-writes to execution-state.json + epic markdown.
Manages git worktrees for parallel task and story execution. Operations: create, list, remove, cleanup, detect-context. Follows Rule 14 (Worktree Lifecycle) naming convention under .claude/worktrees/{identifier}/.
Implements a feature/story/task using TDD (Red-Green-Refactor) workflow. Schema-aware: v1 (legacy) runs the original Double-Loop TDD flow with story-section task extraction; v2 (task-first, EPIC-0038) reads task-TASK-XXXX-YYYY-NNN.md + plan-task-TASK-XXXX-YYYY-NNN.md, honours declared I/O contracts, respects task-implementation-map dependencies, verifies post-conditions via grep/assert, and produces a single atomic commit per task via x-git-commit.
Generates a Double-Loop TDD test plan with TPP-ordered scenarios before implementation. Delegates KP reading to a context-gathering subagent, then produces structured Acceptance Tests (outer loop) and Unit Tests in Transformation Priority Premise order (inner loop).
Cleans local git: prune, remove non-main worktrees, delete branches except main/develop.
(Interna) Gera o plano de arquitetura por story (arch-story-*.md) herdando as decisões de nível feature de architecture-decisions-backend.md. Produz apenas os deltas específicos da story (C4 da fatia, sequence diagrams do fluxo, data model, impacto). Não-chamável pelo usuário.
Infrastructure patterns: Docker multi-stage builds, Kubernetes manifests (cloud-agnostic), security context, 12-Factor App principles, graceful shutdown, resource management, and cloud-native design.
Runs tests with coverage reporting and threshold validation.
Runs tests with coverage reporting and threshold validation.
Git operations: branch creation, atomic commits (Conventional Commits), push, and PR creation. Use for any git workflow task including branching, committing, pushing, creating PRs, or managing version control.
Validates phase transitions for orchestrators under Rule 25. Four modes: --mode pre (assert predecessor phases completed before entering phase N), --mode post (assert all child tasks of phase N are completed AND all expected artifacts exist on disk), --mode wave (post-Batch-B verification of parallel wave completeness: N child TaskUpdate completed + N artifacts exist), --mode final (terminal gate composing with x-internal-epic-integrity-gate). Reads execution-state.json.taskTracking.phaseGateResults and TaskList task state; writes back the gate result. Emits a single-line JSON envelope {passed, mode, skill, phase, expectedTasks, completedTasks, missingTasks, expectedArtifacts, missingArtifacts, wallclockMs, timestamp}. Exit 0 on passed, 12 on failure, 13 on malformed args, 14 on task-resolution timeout. First skill in the x-internal-* convention authored by EPIC-0055; eighth overall (after status-update, report-write, args-normalize, story-load-context, story-build-plan, story-verify, story-resume, epic-build-plan, epic-integrity-gate, epic-branch-ensure, story-report) and the eighth under internal/plan/.
Atomic read-modify-write of execution-state.json with flock-based concurrency, schema validation, and idempotency detection. Substitutes inline Edit-based mutations in orchestrator skills (x-epic-implement, x-story-implement, x-pr-fix-epic) that previously suffered race conditions in --parallel mode. PILOT skill for the x-internal-* convention: internal visibility, non-user-invocable, subdir scoping under internal/ops/.
Kustomize patterns: directory structure, patches, components, secret management, generators, and patch types for environment management.
Complete security reference: OWASP Top 10, security headers, secrets management, input validation, cryptography (TLS, hashing, key management), and pentest readiness checklist. Read during security reviews or when implementing security-sensitive features.
Transactional Outbox Pattern: reliable event publishing with polling publisher and CDC strategies, outbox table design, and anti-patterns.
Reconciles execution-state.json (telemetry) against the **Status:** field of Epic / Story markdown artifacts. Default mode (diagnose) is read-only and prints a divergence table. Opt-in --apply rewrites the markdowns atomically via StatusFieldParser and commits via x-git-commit. Respects Rule 19 (legacy v1 epics skip silently) and Rule 22 (markdown is SoT; state.json is telemetry). Use for manual recovery of legacy epics whose markdown status drifted from execution checkpoints.
Thin orchestrator (~460 lines — story-0049-0018 refactor) that drives an epic end-to-end via 6 delegated phases: Phase 0 (args via x-internal-args-normalize), Phase 1 (load+plan via x-internal-epic-build-plan), Phase 2 (epic branch via x-internal-epic-branch-ensure), Phase 3 (sequential-by-default story loop via x-story-implement), Phase 4 (integrity gate + report via x-internal-epic-integrity-gate + x-internal-report-write), Phase 5 (final PR epic/XXXX → develop via x-git-merge + x-pr-create). Defaults flipped by EPIC-0049: sequential execution (opt-in parallel via --parallel), auto-merge of story PRs into epic/XXXX (target changed from develop). Legacy EPIC-0042 behavior preserved under --legacy-flow (auto-detected via execution-state.json flowVersion=1). Zero inline git/gh/jq/mvn calls — orchestrator uses only Read/Glob + Skill.
Loads a story file, validates predecessor dependencies against execution-state.json, executes artifact pre-checks (mtime-based staleness detection across the 7 planning artifacts), classifies scope (SIMPLE / STANDARD / COMPLEX) from task-count + Gherkin-scenario heuristics, and detects planning mode (PRE_PLANNED / HYBRID / INLINE). Replaces ~140 lines of inline Phase 0 logic previously duplicated inside x-implement-story. Fourth skill in the x-internal-* convention (after x-internal-update-status pilot, x-internal-write-report, and x-internal-normalize-args): internal visibility, non-user-invocable, read-only, subdir scoping under internal/plan/.
Refine a bug report and set Refinement Verdict approved/rejected
Atomic read-modify-write of execution-state.json with flock-based concurrency, schema validation, and idempotency detection. Substitutes inline Edit-based mutations in orchestrator skills (x-epic-implement, x-story-implement, x-pr-fix-epic) that previously suffered race conditions in --parallel mode. PILOT skill for the x-internal-* convention: internal visibility, non-user-invocable, subdir scoping under internal/ops/.
Disaster recovery patterns: DR strategies, RPO/RTO, failover automation, DR testing, multi-region patterns, and recovery procedures
Merge-train automation: discovers, validates, and merges a sequence of PRs into develop in deterministic order. Supports --prs, --epic, and --pattern discovery modes with pre-merge validation and dry-run auditing.
Automates ADR generation from architecture plan mini-ADRs: extracts inline decisions, expands to full ADR format, assigns sequential numbering, updates the ADR index, and adds cross-references.
Generates CHANGELOG.md from Conventional Commits history. Parses git log, groups by commit type, maps to Keep a Changelog sections (Added, Changed, Fixed, etc.), and performs incremental updates preserving existing entries.
Task-level PR creation with formatted title, automatic labels, structured body, and target branch logic. Creates standardized PRs for individual tasks with Task ID traceability.
Cleans local git state in one pass: fetches origin with prune, removes all non-main worktrees (any path), and deletes all local branches except main/master/develop. Destructive by default with an interactive y/N confirmation gate; supports --dry-run (preview) and --yes (non-interactive).
Runs automated smoke tests against the TCP socket server using a standalone Java client with message framing and protocol validation.
Refine a bug report and set Refinement Verdict approved/rejected
Detects the resumable state of an in-flight story: reads plans/epic-XXXX/execution-state.json via x-internal-status-update --read-only, identifies the first PENDING or IN_PROGRESS task (the resume point), catalogues DONE tasks with their commitSha, extracts the last committed SHA, and flags staleness when the story file's mtime is newer than any DONE task's completion timestamp. Emits a single-line JSON envelope {resumePoint, tasksCompleted, tasksPending, lastCommitSha, staleWarnings}. Seventh skill in the x-internal-* convention and the fourth under internal/plan/ (after x-internal-story-load-context, x-internal-story-build-plan, x-internal-story-verify). Read-only by construction — never mutates state.
Generate detailed User Story files from an Epic and system specification with full data contracts, Gherkin acceptance criteria, Mermaid sequence diagrams, dependency declarations, tagged sub-tasks, quality gate validation, and optional Jira integration.
Task-level PR creation with formatted title, labels, structured body, and target branch.
Automates ADR generation from architecture plan mini-ADRs: extracts inline decisions, expands to full ADR format, assigns sequential numbering, updates the ADR index, and adds cross-references.
Validates event schemas, producer/consumer patterns, error handling, dead letter topics, and operational readiness for event-driven architectures.
Complete testing reference: testing philosophy, 8 test categories, coverage thresholds, fixture patterns, data uniqueness, async handling, database strategy, and {{LANGUAGE}}-specific test frameworks. Read before writing tests.
Creates Conventional Commits with Task ID in scope and pre-commit chain (format -> lint -> compile). Central commit point in the task-centric workflow with TDD tag support.
Tech Lead holistic review with {review_max_score}-point checklist covering Clean Code, SOLID, architecture, framework conventions, tests, TDD process, security, and cross-file consistency. Produces GO/NO-GO decision. Use for final review before merge.
Generate an Epic document from spec analysis: cross-cutting rules, story index, DoR/DoD, optional Jira. Invoked only by x-create-feature (Phase 2). Not user-invocable.
Atomic read-modify-write of execution-state.json with flock-based concurrency, schema validation, and idempotency detection. Substitutes inline Edit-based mutations in orchestrator skills (x-implement-epic, x-implement-story, x-fix-epic-pr) that previously suffered race conditions in --parallel mode. PILOT skill for the x-internal-* convention: internal visibility, non-user-invocable, subdir scoping under internal/ops/.
Parses an argv string against a declarative JSON schema with typed flags (boolean, string, integer, enum), defaults, mutually-exclusive groups, and deprecation warnings, then emits a normalized `{parsed, warnings, errors}` envelope on stdout. Replaces ~150 lines of inline argv parsing inlined inside `x-implement-epic`, `x-implement-story`, and `x-orchestrate-epic`, giving every orchestrator identical flag-validation syntax and error messages. Third skill in the x-internal-* convention (after x-internal-update-status pilot and x-internal-write-report): internal visibility, non-user-invocable, subdir scoping under internal/ops/.
Hybrid changelog: narrative Highlights + Keep-a-Changelog sections from v2 features.
Discovers all PRs from an epic via execution-state.json, fetches and classifies review comments in batch, generates a consolidated findings report, applies fixes, and creates a single correction PR. Supports dry-run, explicit PR list fallback, and idempotent re-execution.
Generate detailed User Story files from an Epic and system specification with full data contracts, Gherkin acceptance criteria, Mermaid sequence diagrams, dependency declarations, tagged sub-tasks, quality gate validation, and optional Jira integration.
Assists migration of a v1 epic document to the v2 value-driven template (EPIC-0070). Parses v1 technical blocks (Packages, Contratos, SOLID, Observabilidade), classifies each block with a safe default heuristic, and optionally asks the operator for confirmation per block (--interactive). Side-effects: writes epic.md in v2 format atomically, creates ADRs for 'virar ADR' decisions, updates system.md via x-update-system-architecture. Supports --dry-run and recovery from interrupted sessions.
Create a structured bug report from the RA9 template
Runs tests with coverage reporting and threshold validation. Use whenever writing, running, or analyzing tests. Triggers on: test, coverage, TDD, unit test, integration test, test failure, coverage gap, or Definition of Done validation.
Create a complete feature (Epic + N Stories + Implementation Map) from a spec file, with a worktree-isolated docs/ branch, consolidated commit, and auto-merged PR into epic/XXXX.
Implements a bug story (regression-test or fix) end-to-end via BUG_GATE, ordering enforcement, context loading, and delegation to x-internal-implement-story.
Create a complete feature (Epic + N Stories + Implementation Map) from a spec file, with a worktree-isolated docs/ branch, consolidated commit, and auto-merged PR into epic/XXXX.
Dynamic Application Security Testing -- tests the running application for XSS, injection, misconfiguration, and information disclosure using OWASP ZAP or Nuclei.
Classifies the git working tree state (CLEAN/DIRTY/DIVERGENT/AMBIGUOUS) and returns a stable exit code. Invoked by orchestrators before starting story/task execution to detect dirty or divergent states early. Exit 0 = CLEAN or DIRTY; exit 15 (WORKTREE_AMBIGUOUS) = DIVERGENT or AMBIGUOUS (unless --allow-dirty).
Build gate verification between parallelism groups. Compiles code, classifies errors, decides retry vs escalate, extracts outputs for next group. Used between each implementation group in Phase 2.
CI/CD pipeline patterns: build stages, test parallelization, security scanning, artifact management, environment promotion, approval gates, caching strategies, and language-specific pipeline configurations.
Sets up the local development environment including container orchestrator, database, and build tools.
Engine compartilhada de implementação de story (Phases 0.3→3). Usada por x-implement-story e x-fix-bug.