tunneleven
18 verified skills
c4flow:code
Execute code implementation via a strict serial task loop — pick one unblocked task, TDD it, verify, review, PR, merge, then pick the next task. Each task gets its own branch (feat/bead-id-slug). Uses bd ready → atomic claim → set in-progress → dolt push → TDD (RED gate pause) → verify → review → PR/merge → close → repeat. Trigger when the user wants to start coding, implement tasks, or run the implementation phase.
development
c4flow:init
Initialize C4Flow dependencies in the current project — auto-sets up git (init repo if missing), installs Dolt and Beads. Use when setting up a new project, when bd/dolt is missing, or when the user needs to configure DoltHub sync or team collaboration.
testing
c4flow:review
AI code review via Codex subagent — writes quality-gate-status.json, resolves beads gate on pass. Use when the user wants code reviewed, needs a quality check before PR, or mentions "review", "code review", or "codex review". Also triggers before creating pull requests.
development
c4flow:research
Perform merged market analysis and technical research on a feature idea. Combines market-research (business domain, competitive landscape, feature comparison, gap analysis) with C4Flow technical research (approaches, contrarian view, risks).
data-ai
c4flow:pr
Create a pull request with spec summary, task list from beads epic, and test/review results. Use when the user wants to create a PR, open a pull request, or submit code for merge. Triggers on "create PR", "open pull request", "submit for review", or after verify passes.
development
c4flow:merge
[NOT IMPLEMENTED] Merge PR to main — pre-checks, merge strategy, cleanup. Tell the user this skill is not yet implemented.
testing
c4flow:infra
Provision AWS EC2 + nginx + SSL infrastructure and Cloudflare DNS for the current C4Flow feature. Use when the user runs /c4flow:infra or when the orchestrator advances to INFRA state.
devops
c4flow
Orchestrate the complete c4flow agentic development workflow — from research through deployment. Use when the user mentions "c4flow", wants to start a new feature workflow, or asks about the development pipeline. Triggers on feature planning, implementation orchestration, and workflow management.
development
c4flow:deploy
Set up GitHub Actions CI/CD and deploy the application to the EC2 infrastructure provisioned by c4flow:infra. Use when the user runs /c4flow:deploy or when the orchestrator advances to DEPLOY state. Requires infra to be provisioned first.
testing
c4flow:beads
Break down a feature into a detailed beads epic with rich task metadata — priority, assignee, acceptance criteria, design notes, dependency graph (blocks/waits-for/conditional), and spec links. Use when the user wants to plan work, decompose features into tasks, create an epic, or set up a dependency-aware work graph. Triggers on task breakdown, epic creation, work planning, or "break this into tasks".
data-ai
c4flow:design
Generate design system and UI mockups for a feature using Pencil MCP. Runs after SPEC phase, before BEADS. Produces MASTER.md (design tokens), screen-map.md (screen breakdown), and a .pen file with reusable components and screen frames. Use when the workflow reaches DESIGN state or user asks to design screens/UI. Integrates impeccable best practices OKLCH colors, tinted neutrals, modular typography, spatial rhythm, motion tokens, 8-state component design, and UX writing.
tools
c4flow:e2e
Run end-to-end browser tests against the deployed or locally-running application. Auto-detect e2e framework (Playwright, Cypress, Selenium, TestCafe, Robot Framework), manage app server lifecycle, execute tests, classify failures, and report results. Manual trigger — not part of auto-flow. Use when the user wants to run e2e tests, browser tests, integration tests against a live app, or validate user flows. Triggers on "e2e", "end-to-end", "browser test", "playwright", "cypress", or "smoke test".
development
c4flow:spec
Generate structured spec artifacts (proposal, tech-stack, spec, design) through interactive collaboration, using research.md as structured input.
content-media
c4flow:tdd
Test-driven development — RED-GREEN-REFACTOR cycles for all C4Flow implementation work. Merged into c4flow:code as a sub-agent phase with a mandatory RED gate pause. Use c4flow:code to run the full task loop.
development
c4flow:infra-destroy
Tear down all AWS infrastructure provisioned by c4flow:infra — destroys EC2, VPC, Elastic IP, Cloudflare DNS record, and removes associated GitHub Secrets. Clears infraState from .state.json. Requires explicit double-confirmation before any destructive action. Use when the user runs /c4flow:infra-destroy or asks to "destroy infra", "tear down", or "clean up AWS resources".
testing
c4flow:sync
Sync local project with remote sources — pulls DoltHub beads and GitHub repo to local. Handles the "no common ancestor" Dolt error that occurs when bd init creates a fresh local DB that conflicts with an existing DoltHub history. Use when local beads are out of sync, after a fresh init on a project that already has DoltHub data, or to pull the latest GitHub changes.
testing
c4flow:test
Run unit and integration tests with coverage checking. Auto-detect framework, classify failures, enforce coverage threshold before advancing to review. Use when the user wants to run tests, check coverage, or validate implementation quality. Triggers on "run tests", "check coverage", "test suite", or when the code phase completes.
development
c4flow:verify
Quality gate aggregation — runs bd preflight, combines with Codex review results, declares Ready for PR status. Use when the user wants to check if code is ready for PR, verify quality gates, or run preflight checks. Also triggers when mentioning "verify", "preflight", "quality gate", or "ready for PR".
development