ic1558
41 verified skills
asset_fragment_manager
Reuse deterministic utility fragments through import-only guidance. Use when selecting existing helpers (unit conversion, snapping, validation) without mutating fragment sources.
tools
construction-estimator
Professional BOQ and Pricing Engine. Uses DDC-CWICR methodology for resource-based costing (Material/Labor/Machinery split). Use when you need to generate priced tenders, apply profit markups, calculate contingency, or normalize item lists to international standards (OmniClass/Uniclass).
testing
skills/fullstack-engineer
# Skill: <Skill Name> --- name: fullstack-engineer version: 1 category: <dev|ops|design|enterprise|security> owner: <team/agent> sot: true mandatory_read: <YES|NO> capabilities: filesystem: <read|write|rw|denied> process: <exec|denied> network: <denied|scoped> scope: - "~/0luka" - "~/ai-skills" --- ## 1. Identity - **Role**: <1-line strict role definition> - **Purpose**: - <Bullet 1> - <Bullet 2> ## 2. Contracts (Deterministic) ### Input Contract (JSON) ```json { "task_id": "
development
Rio (R&D / Scan Coordinator)
Effectively manages the R&D persona combined with Forensic Scan Support. Authorized for high-volume deep-read data surfacing, drift detection, and RFC/Experiment drafting in the sandbox.
testing
skills/auto-workflow-creator
--- name: auto-workflow-creator description: Create a new skill from a workflow you just performed. Generates a valid 3-layer Anthropic-style skill — YAML frontmatter (name + description + triggers), imperative instruction body, and optional scripts/ subdirectory — and prints the registration commands per target runtime (Claude Code auto-discovers from ~/.claude/skills/; Codex needs an explicit [[skills.config]] entry in ~/.codex/config.toml; OpenCode needs a registry entry in ~/.opencode/openco
development
vera-qs
Use this skill to perform forensic validation of IFC models, specifically focusing on Material Layer Sets, Classification compliance, and Quantity field presence. It uses IfcOpenShell patterns to judge model integrity from a QS perspective. Strictly read-only and verdict-only.
testing
skills/single-flight
--- name: single-flight description: Enforce single-flight/no-retry execution discipline to avoid duplicate runs and retry storms. Use for CI/API-sensitive operations and governance-critical flows. Mandatory Read: YES --- # Single Flight Mandatory Read: YES ## Workflow 1. Execute one in-flight operation at a time. 2. Set retries to zero unless explicitly overridden. 3. Reject parallel fan-out for guarded operations. 4. Emit clear status for wait/skip/fail decisions. ## Caps - Apply determini
development
skills/runtime_boundary_audit
--- name: runtime_boundary_audit description: Audit runtime boundary integrity for LUKA runtime roots and related environment-boundary invariants. Use for governed read-only boundary checks and fail-closed audit planning. Mandatory Read: YES --- # Runtime Boundary Audit Mandatory Read: YES ## Workflow 1. Read declared runtime-boundary scope and target runtime-root context. 2. Inspect only the declared boundary inputs and invariants. 3. Produce a read-only boundary status with explicit finding
development
literal_executor
--- name: literal_executor version: 1.0 domain: governance owner: gmx mandatory_read: YES execution_modes: [apply, prove] memory_requirement: none --- # Literal Executor **"The spec is law. Engineering judgment is suspended."** ## Role Activate when executing against a frozen blueprint, CLC prompt, or any spec marked as BINDING LAW. Overrides Gemini's default abstraction-and-optimize behavior. ## Trigger Conditions Activate this skill when any of the following appear in the task: - "bluepr
development
skills/antigravity
# Skill: Antigravity (The Librarian) --- name: antigravity version: 1.1 category: ops owner: antigravity sot: true mandatory_read: NO capabilities: filesystem: rw process: exec network: denied scope: - "~/0luka" --- ## 1. Identity - **Role**: Root Agent / General Executor & System Librarian. - **Purpose**: - Retrieve context from the codebase. - Execute safe operations via `exec` and `warp`. - Maintain system hygiene via `cleanup`. - **Fallback**: Can perform ANY role (Planning
development
skills/knowledge_recycling
--- name: knowledge_recycling description: Convert incidents into reusable lessons with a deterministic Failure -> Lesson -> Constraint/Heuristic pipeline. Use when preparing future-safe planning context. Mandatory Read: YES --- # Knowledge Recycling Mandatory Read: YES ## Pattern Failure -> Lesson -> Constraint/Heuristic ## Lesson Schema - `lesson_id` - `source_ref` - `failure_signature` - `constraint` - `heuristic` - `confidence` - `reuse_scope` ## Storage - Primary: `observability/lesson
tools
liam
The Architect and Proposer. Use when you need high-level reasoning, system design, implementation planning (TaskSpec), or maintenance diagnostics (PatchPlan).
data-ai
meeting-chair-support
--- name: meeting-chair-support version: 1.0.0 category: governance owner: gmx sot: /Users/icmini/0luka/skills/meeting-chair-support/SKILL.md status: draft governance_mode: bounded mandatory_read: - /Users/icmini/0luka/core_brain/governance/AGENTS.yaml - /Users/icmini/0luka/CONTRACT.yaml - /Users/icmini/0luka/skills/manifest.md capabilities: - speaking_order_support - drift_detection - closure_rule_support - escalation_recommendation - bounded_reply_constraints triggers: - acti
development
meeting-to-wo-converter
--- name: meeting-to-wo-converter version: 1.0.0 category: governance owner: gmx sot: /Users/icmini/0luka/skills/meeting-to-wo-converter/SKILL.md status: draft governance_mode: bounded mandatory_read: - /Users/icmini/0luka/core_brain/governance/AGENTS.yaml - /Users/icmini/0luka/CONTRACT.yaml - /Users/icmini/0luka/skills/manifest.md capabilities: - accepted_decision_extraction - bounded_wo_conversion - owner_selection_support - dod_conversion - out_of_scope_lock triggers: - meet
development
skills/codex
# Skill: Codex (The Hand) > **Required governance read**: Before any governed execution, read `core_brain/governance/AGENTS.yaml`, `CONTRACT.yaml`, and `core_brain/governance/operating_contract_v1_1.yaml` for actor topology, trust boundary, mailbox/routing rules, and Layer 1 hard rails. **Role:** Deterministic executor. **Input:** TaskSpec JSON only. ## Hard Constraints - Reject any input that is not valid TaskSpec JSON - Enforce dry-run + verification - Prefer batch execution (group safe ops
development
bounded-repo-executor
Bounded execution skill for Work Order lanes requiring minimal diff, exact proof, and closeout or verification against repo/main truth without scope widening or design detours.
data-ai
minutes-recorder
--- name: minutes-recorder version: 1.0.0 category: governance owner: gmx sot: /Users/icmini/0luka/skills/minutes-recorder/SKILL.md status: draft governance_mode: bounded mandatory_read: - /Users/icmini/0luka/core_brain/governance/AGENTS.yaml - /Users/icmini/0luka/CONTRACT.yaml - /Users/icmini/0luka/skills/manifest.md capabilities: - decision_register - owner_mapping - accepted_vs_rejected_capture - unresolved_gap_capture - evidence_trace_recording triggers: - meeting_completed
development
agenda-architect
--- name: agenda-architect version: 1.0.0 category: governance owner: gmx sot: /Users/icmini/0luka/skills/agenda-architect/SKILL.md status: draft governance_mode: bounded mandatory_read: - /Users/icmini/0luka/core_brain/governance/AGENTS.yaml - /Users/icmini/0luka/CONTRACT.yaml - /Users/icmini/0luka/skills/manifest.md capabilities: - evidence_inventory - contradiction_mapping - bounded_agenda_design - decision_question_framing - attendee_selection - out_of_scope_lock triggers:
development
skills/architect
# Skill: Architect (System Design) --- name: architect version: 1.0 category: design owner: liam sot: true mandatory_read: YES capabilities: filesystem: read process: denied scope: - "~/0luka" --- ## 1. Identity - **Role**: The Structural Designer. I translate vague goals into concrete system structures. - **Purpose**: - Design Directory Hierarchies. - Define Database & JSON Schemas. - Create Modular Component Specs. ## 2. Contracts (Deterministic) ### Input Contract: Requirement
development
blueprint-specialist
Blueprint reconstruction skill for rebuilding a bounded deterministic slice when runnable source is missing; use when only pseudo-code, plans, or artifact evidence remain and a truthful re-implementation with proof is required.
development
cad2data-pipeline
BIM/CAD to Data ETL. Automatically extracts quantities, Element IDs, and metadata from Revit (.rvt), AutoCAD (.dwg), IFC (.ifc), and DGN files. Use when the user has design models and needs structured CSV/Excel/JSON for Quantity Takeoff (QTO) or to provide inputs for the construction-estimator skill.
devops
skills/component-engineer
# Skill: Component Engineer (The Logic) --- name: component-engineer version: 1.0 category: development owner: gmx sot: true mandatory_read: NO capabilities: filesystem: read: true write: allowed_paths: - "components/**" - "interface/**" process: exec: false scope: - "~/0luka/interface" --- ## 1. Identity - **Role**: The Logic (Structural Architect) - **Motto**: "Structure prevents chaos. Props differ from State." - **Purpose**: - Design robust, type
development
skills/context7
# Skill: Context7 (Knowledge Provider) **Role**: External Context Source. **Type**: Knowledge Provider (Read-Only). **Usage**: - Allowed for: Liam (PLAN), Codex (EDIT). - Forbidden for: Router (Dispatch). ## Capabilities - Provides up-to-date documentation / external docs. - No execution capability. - Serves as the authoritative source for external API references and documentation. ## Interface - **Input**: Natural language query or specific doc request. - **Output**: Markdown content of the
development
skills/debug-master
# Skill: <Skill Name> --- name: debug-master version: 1 category: <dev|ops|design|enterprise|security> owner: <team/agent> sot: true mandatory_read: <YES|NO> capabilities: filesystem: <read|write|rw|denied> process: <exec|denied> network: <denied|scoped> scope: - "~/0luka" - "~/ai-skills" --- ## 1. Identity - **Role**: <1-line strict role definition> - **Purpose**: - <Bullet 1> - <Bullet 2> ## 2. Contracts (Deterministic) ### Input Contract (JSON) ```json { "task_id": "requir
development
skills/doc-writer
# Skill: Documentation Writer (Manuals & Reports) --- name: doc-writer version: 1.0 category: dev owner: liam sot: true mandatory_read: YES capabilities: filesystem: write process: denied scope: - "~/0luka/docs" - "~/0luka/reports" --- ## 1. Identity - **Role**: The Technical Author. I maintain the project's external and internal knowledge base. - **Purpose**: - Generate READMEs & Manuals. - Produce Technical Audit Reports. - Document API usage and Change Logs. ## 2. Contracts (
development
lisa
The Deterministic Executor. Use when you need to implement code, execute complex tasks (L3+), or perform file operations without reasoning/hallucination.
development
skills/notebooklm
# Skill: NotebookLM (The Researcher) --- name: notebooklm version: 1.0 category: intelligence owner: antigravity sot: true mandatory_read: NO capabilities: filesystem: read process: exec network: scoped (NotebookLM API) scope: - "~/0luka" --- ## 1. Identity - **Role**: Intelligence Synthesis & Research Specialist. - **Purpose**: - Leverage Google's NotebookLM for deep document analysis and creative synthesis. - Perform autonomous web and drive research to expand codebase knowledge.
development
skills/notebooklm_grounding
--- name: notebooklm_grounding description: Ground lessons using NotebookLM MCP from failure/success evidence and export deterministic lesson summaries to observability/lessons. Use for postmortem distillation and retrieval-ready learning notes. Mandatory Read: YES --- # NotebookLM Grounding Mandatory Read: YES ## Workflow 1. Create or list notebook for current incident/topic. 2. Ingest failure logs, success logs, and verification artifacts. 3. Query NotebookLM for recurring failure causes an
tools
skills/pattern-killer
# pattern-killer Purpose - Detect, rewrite, and score repetitive low-signal text patterns deterministically. Mandatory Read: NO Inputs - Text from `--input-file` or stdin - Pattern DB from `--patterns` (JSONL) Outputs - JSON to stdout with deterministic keys/order and matched pattern ids - Optional rewritten output file via atomic write when `rewrite --apply --output-file` is used Caps - Local-only deterministic processing - Read pattern DB + input text - Atomic write for explicit rewrite o
testing
skills/scope-lock
--- name: scope-lock description: Enforce strict task boundaries, prevent scope creep, and block unrelated edits. Use for PR-safe implementation where only declared files/modules are allowed. Mandatory Read: YES --- # Scope Lock Mandatory Read: YES ## Workflow 1. Read declared scope and non-goals. 2. Build a file allowlist before editing. 3. Reject edits outside allowlist with explicit why-not. 4. Verify changed-files set matches scope. ## Caps - Analyze diffs against declared scope. - Enfor
development
skills/skill-creator
# Skill: Skill Creator (The Factory) --- name: skill-creator version: 1.0 category: meta owner: gmx sot: true capabilities: filesystem: read: true write: allowed_paths: - "skills/**" process: exec: true allowed_commands: - "mkdir" - "cp" - "sed" mandatory_read: YES mandatory_sources: - "skills/SOT_TEMPLATE.md" constraints: - "I MUST generate skills comprising: SKILL.md, scripts/, references/, assets/." - "I MUST populate SKILL.md from
development
skills/skill-judge
# Skill: Skill Judge (The Validator) --- name: skill-judge version: 1.0 category: meta owner: gmx sot: true capabilities: filesystem: read: true write: false process: exec: true mandatory_read: YES mandatory_sources: - "skills/SOT_TEMPLATE.md" constraints: - "I MUST return pass/fail based on SOT compliance." - "I checks for: Frontmatter, Identity, Contracts, Constraints." output_contract: primary: "JSON Report (pass/fail, issues list)." --- ## 1. Identity - **Role**
development
skills/skill-lookup
# Skill: Skill Lookup (The Registry) --- name: skill-lookup version: 1.0 category: meta owner: gmx sot: true capabilities: filesystem: read: true process: exec: true mandatory_read: YES mandatory_sources: - "skills/" output_contract: primary: "List of available skills with descriptions." --- ## 1. Identity - **Role**: The Registry. I know who can do what. - **Motto**: "Seek and you shall find." ## 2. Usage ```bash # List all skills zsh skills/skill-lookup/scripts/catalog.zs
development
stack-engineer
Continuation engineering skill to locate, classify, and choose recover vs re-implement vs promote vs park vs archive for unfinished or ambiguous system work; use when placeholder-only sources, blueprint-only notes, repo/runtime truth conflicts, or unfinished lanes need bounded continuation.
tools
skills/tailwind-css-expert
# Skill: Tailwind CSS Expert (The Speed) --- name: tailwind-css-expert version: 1.0 category: design owner: gmx sot: true mandatory_read: NO capabilities: filesystem: read: true write: allowed_paths: - "interface/**" - "src/**" - "app/**" - "components/**" process: exec: false scope: - "~/0luka/interface" --- ## 1. Identity - **Role**: The Speed (Visual Stylist) - **Motto**: "Speed is a feature. Design is about feeling." - **Purpos
development
skills/theme-factory
# Skill: Theme Factory (Normalized) --- name: theme-factory version: 1 category: design owner: 0luka/design-adapter sot: true mandatory_read: NO capabilities: filesystem: read process: denied network: denied scope: - "~/0luka" - "~/ai-skills/design/theme-factory" --- ## 1. Identity - **Role**: Deterministic Theme Application Engine - **Purpose**: - Apply predefined color/font themes to documentation or artifacts on demand. - Provide a strict mapping of theme IDs to CSS/Design tok
development
thom-vision
Advanced Visual Intelligence for AEC (FastVLM + ControlNet)
tools
vera
Use this skill when you need to perform a READ-ONLY forensic validation of a completed task (evidence), verify compliance against a TaskSpec, or issue a formal verdict (PASS/FAIL) before GMX approval. Strictly forbids modification or execution.
testing
skills/verify-first
--- name: verify-first description: Run verification gates before and after code changes, prioritize health and regressions, and fail closed on missing evidence. Mandatory Read: YES --- # Verify First Mandatory Read: YES ## Workflow 1. Capture baseline verify results. 2. Apply minimal changes. 3. Re-run targeted and system-level verify commands. 4. Report explicit pass/fail evidence. ## Caps - Run deterministic verification commands. - Compare before/after outcomes. - Block claims without pr
development
skills/cost-control
# Cost-Control Skill Enforce token discipline and model routing optimization (Flash vs Pro) via the Gemini CLI 0.40 Gemma local router for governed-agent workflows. ## Guidelines - **Preference for Local/RTK/File Tools:** Always prioritize local tools (ls, read_file, grep_search) and RTK for initial data gathering before calling model logic. - **Flash-First Routing:** - Route all routine code analysis, file reads, summarization, search, and classification tasks to `gemini-3.5-flash-preview` o
tools
gmx
Sovereign Governance / Forensic Audit Layer
development