skills.md

Flow-Code Skills Reference

Skills are organized into core (essential workflow) and extensions (optional capabilities).

Core Skills (7)

These skills form the primary workflow. They ship with the plugin.

Skill	Command	Purpose	Location
flow-code-go	/flow-code:go	Full autopilot — brainstorm → plan → work → review → close → PR (sole user-facing entry point)	skills/
flow-code-run	Internal	Internal pipeline engine, invoked by /flow-code:go	skills/
flow-code-plan	/flow-code:plan	Research codebase and create epic with tasks	codex/skills/
flow-code-work	/flow-code:work	Execute tasks with re-anchoring, reviews, wave checkpoints	codex/skills/
flow-code	/flow-code	Task/epic management entry point (list, create, status)	skills/
flow-code-setup	/flow-code:setup	Install flowctl CLI and configure project	skills/
flow-code-map	/flow-code:map	Generate codebase architecture maps	skills/

Extension Skills (47)

Optional capabilities that extend the core workflow. Skills live in skills/ and/or codex/skills/ (54 unique skills total, 7 core + 47 extension).

Review & Quality

Skill	Command	Purpose
flow-code-plan-review	Internal	Cross-model plan validation (RP or Codex)
flow-code-impl-review	Internal	Per-task implementation review with fix loop
flow-code-epic-review	Internal	Epic completion adversarial gate
flow-code-code-review	Internal	Five-axis scoring with severity labels (Worker Phase 6, impl-review, PR review)

Development Extensions

Skill	Command	Purpose
flow-code-brainstorm	/flow-code:brainstorm	Explore and pressure-test ideas (interactive or --auto self-interview)
flow-code-debug	/flow-code:debug	Systematic debugging with root cause investigation
flow-code-auto-improve	/flow-code:auto-improve	Autonomous code quality improvement loops
flow-code-django	/flow-code:django	Django-specific patterns, security, and testing
flow-code-deps	/flow-code:deps	Dependency graph visualization and execution order
flow-code-api-design	/flow-code:api-design	API design and module boundary review
flow-code-performance	/flow-code:performance	Performance investigation, optimization, and benchmarks
flow-code-tdd	/flow-code:tdd	Test-first methodology, Prove-It Pattern, coverage strategy
flow-code-incremental	Internal	Vertical slicing, incremental commits, Implement-Test-Verify-Commit cycle
flow-code-simplify	/flow-code:simplify	Reduce code complexity while preserving exact behavior
flow-code-frontend-ui	/flow-code:frontend-ui	Production-quality UI: components, layouts, state, accessibility

Architecture & Infrastructure Skills

Skill	Command	Purpose
flow-code-security	Internal	OWASP Top 10, three-tier security boundaries
flow-code-auth	Internal	Authentication, authorization, OAuth, JWT, RBAC, session management
flow-code-database	Internal	Schema design, migrations, query optimization, ORM patterns
flow-code-caching	Internal	HTTP cache, CDN, Redis, in-memory, cache invalidation strategy
flow-code-containerization	Internal	Dockerfiles, docker-compose, Kubernetes, image optimization
flow-code-microservices	Internal	Service boundaries, inter-service communication, saga, event-driven
flow-code-realtime	Internal	WebSocket, SSE, long-polling, connection management, scaling
flow-code-state-management	Internal	Frontend/full-stack state architecture, tool selection, patterns
flow-code-error-handling	Internal	Error classification, retry logic, circuit breakers, graceful degradation
flow-code-i18n	Internal	Multi-language support, locale formatting, RTL layouts
flow-code-monitoring	Internal	Dashboards, SLOs/SLIs, alerting rules, on-call runbooks
flow-code-observability	Internal	Logging, tracing, metrics, health endpoints
flow-code-documentation	Internal	ADRs, API docs, READMEs, changelogs, doc-as-code workflow

Workflow Extensions

Skill	Command	Purpose
flow-code-interview	/flow-code:interview	Refine specs through structured Q&A (40+ questions)
flow-code-sync	/flow-code:sync	Sync downstream task specs after implementation drift
flow-code-retro	/flow-code:retro	Post-epic retrospective and lessons learned
flow-code-prime	/flow-code:prime	Assess codebase readiness for agent work (8 pillars, 48 criteria)
flow-code-autoplan	/flow-code:autoplan	Multi-perspective auto-review pipeline (CEO, eng, design, DX)
flow-code-qa	/flow-code:qa	Visual QA testing with browser automation
flow-code-design-review	/flow-code:design-review	Visual design audit with browser automation
flow-code-guide	Internal	Skill discovery flowchart — helps select the right skill

Tooling Extensions

Skill	Command	Purpose
flow-code-ralph-init	/flow-code:ralph-init	Scaffold autonomous Ralph harness
flow-code-loop-status	/flow-code:loop-status	Monitor running Ralph/auto-improve loops
flow-code-worktree-kit	/flow-code:worktree-kit	Git worktree management for parallel work
flow-code-export-context	/flow-code:export-context	Export context for external model review
flow-code-rp-explorer	/flow-code:rp-explorer	RepoPrompt-powered codebase exploration
flow-code-skill-create	/flow-code:skill-create	Create new flow-code skills
flow-code-prompt-eng	Internal	Prompt engineering guidance for review agents
flow-code-cicd	/flow-code:cicd	CI/CD pipeline setup, quality gates, and deployment automation
flow-code-context-eng	/flow-code:context-eng	Context window management and optimization
flow-code-deprecation	/flow-code:deprecation	Feature, API, and module deprecation workflows
browser	/browser	Browser automation via agent-browser CLI

Recommended Usage Order

For a new project:

1. /flow-code:setup — install and configure
2. /flow-code:prime — assess codebase readiness
3. /flow-code:go "idea" — full autopilot: brainstorm → plan → work → review → close → PR

Or for more control:

1. /flow-code:brainstorm --auto "idea" — AI self-interview, produces requirements doc
2. /flow-code:plan .flow/specs/<slug>-requirements.md — research + create tasks
3. /flow-code:work fn-N — execute tasks

/flow-code:go fn-N resumes an existing epic (brainstorm auto-skipped).

Tier Classification

Skills are classified into four tiers based on complexity and decision-making authority, inspired by the gstack tiering system. The tier field in each skill's YAML frontmatter enables tooling to filter, prioritize, and display skills by capability level.

Tier 1 — Basic Tools

Simple utilities and wrappers that perform a single focused action.

Skill	Purpose
flow-code-setup	Install flowctl CLI and configure project
flow-code-worktree-kit	Git worktree management
flow-code-export-context	Export context for external model review
flow-code-deps	Dependency graph visualization
flow-code-loop-status	Monitor running loops
flow-code-guide	Skill discovery flowchart
browser	Browser automation via agent-browser CLI
flow-code-rp-explorer	RepoPrompt-powered codebase exploration

Tier 2 — Observation & Monitoring

Skills that gather information, analyze codebases, or provide monitoring and guidance.

Skill	Purpose
flow-code-map	Generate codebase architecture maps
flow-code-prime	Assess codebase readiness for agent work
flow-code-context-eng	Context window management and optimization
flow-code-performance	Performance investigation and benchmarks
flow-code-prompt-eng	Prompt engineering guidance for review agents
flow-code-debug	Systematic debugging with root cause investigation
flow-code-monitoring	Dashboards, SLOs/SLIs, alerting rules
flow-code-observability	Logging, tracing, metrics, health endpoints
flow-code-documentation	ADRs, API docs, changelogs, doc-as-code

Tier 3 — Interactive Decision-Making

Skills that involve planning, design, or interactive multi-step workflows.

Skill	Purpose
flow-code-go	Full autopilot (brainstorm → plan → work → review → close)
flow-code-run	Unified pipeline entry point (plan, review, work, close)
flow-code-plan	Research codebase and create epic with tasks
flow-code-work	Execute tasks with re-anchoring and wave checkpoints
flow-code-brainstorm	Explore and pressure-test ideas (interactive or --auto)
flow-code-interview	Refine specs through structured Q&A
flow-code-api-design	API design and module boundary review
flow-code-cicd	CI/CD pipeline setup and deployment automation
flow-code-django	Django-specific patterns, security, and testing
flow-code-skill-create	Create new flow-code skills
flow-code-auto-improve	Autonomous code quality improvement loops
flow-code-sync	Sync downstream task specs after implementation drift
flow-code-ralph-init	Scaffold autonomous Ralph harness
flow-code-deprecation	Feature and API deprecation workflows
flow-code-plan-review	Cross-model plan validation
flow-code-impl-review	Per-task implementation review
flow-code-epic-review	Epic completion adversarial gate
flow-code-code-review	Five-axis scoring with severity labels
flow-code-qa	Visual QA testing with browser
flow-code-design-review	Visual design audit with browser
flow-code-autoplan	Multi-perspective auto-review pipeline
flow-code-tdd	Test-first methodology, Prove-It Pattern
flow-code-incremental	Vertical slicing, incremental commits
flow-code-simplify	Reduce complexity while preserving behavior
flow-code-frontend-ui	Production-quality UI components and layouts
flow-code-security	OWASP Top 10, security boundaries
flow-code-auth	Authentication, authorization, OAuth, JWT, RBAC
flow-code-database	Schema design, migrations, query optimization
flow-code-caching	HTTP cache, CDN, Redis, cache invalidation
flow-code-containerization	Docker, Kubernetes, image optimization
flow-code-microservices	Service boundaries, saga, event-driven patterns
flow-code-realtime	WebSocket, SSE, connection management
flow-code-state-management	Frontend/full-stack state architecture
flow-code-error-handling	Error classification, retry, circuit breakers
flow-code-i18n	Multi-language support, locale formatting, RTL

Tier 4 — Final Decisions & Execution

Skills that make final judgments, complete workflows, or serve as primary management entry points.

Skill	Purpose
flow-code	Task and epic management entry point
flow-code-retro	Post-epic retrospective and lessons learned
flow-code-autoplan	Multi-perspective auto-review pipeline

Template Generation

Skills can be generated from .tmpl template files with {{PLACEHOLDER}} markers that resolve to shared content. This keeps common patterns (preamble, flowctl path, review protocols) in sync across all skills.

Creating a Template

1. Copy the existing SKILL.md to SKILL.md.tmpl in the same directory
2. Replace shared content with placeholder markers (see table below)
3. Add {{GENERATED_NOTICE}} after the frontmatter to mark the file as auto-generated
4. Run the generation script to produce the SKILL.md

Available Placeholders

Placeholder	Resolves To
{{GENERATED_NOTICE}}	<!-- AUTO-GENERATED from SKILL.md.tmpl — DO NOT EDIT DIRECTLY -->
{{FLOWCTL_PATH}}	FLOWCTL="$HOME/.flow/bin/flowctl"
{{SKILL_NAME}}	Extracted from the template's frontmatter name: field
{{PREAMBLE}}	Contents of skills/_shared/preamble.md
{{RP_REVIEW_PROTOCOL}}	Contents of skills/_shared/rp-review-protocol.md

Running Generation

# Generate all SKILL.md files from their .tmpl sources
bash scripts/gen-skill-docs.sh

# Preview what would change without writing
bash scripts/gen-skill-docs.sh --dry-run

# Check if generated files are up to date (for CI)
bash scripts/gen-skill-docs.sh --check

CI Freshness Check

Add to your CI pipeline to ensure generated files stay in sync:

bash scripts/check-skill-freshness.sh

This exits with code 1 if any SKILL.md does not match its .tmpl source, preventing stale generated files from being committed.