fb-claude-skills

Opt the current project into down-tier model delegation by installing a standalone .claude/rules/model-delegation.md rule, optionally with pre-shaped .claude/agents/ definitions (fast-executor, task-coder) and an optional agent-state feedback layer. The rule tells Claude to route well-specified data and coding tasks to a cheaper model in a subagent, keeping judgment-heavy work in the main loop. Use when the user says "set up model routing", "set up model delegation", "use cheaper models for subagents", "delegate to a lower-power model", or wants to remove or update that rule.

Pre-share scan for leaked secrets and privacy-sensitive content. Wraps simonw/scan-for-secrets for literal matching (with JSON/URL/HTML/backslash/unicode escape variants) and composes a ripgrep regex pass for shape-based leaks (other users' home paths, emails, IPv4, MAC, JWTs, API keys). Use when user says "scan for secrets", "check for leaked credentials", "pre-share scan", "redact home paths", "PII scan", "strip my username from transcripts", "scan logs before publishing", "check before sharing", "audit before commit", "find leaked API keys", "scan agent transcript".

Convert JavaScript and TypeScript projects to use bun instead of npm/yarn/pnpm. Use when user asks to "migrate from npm to bun", "convert to bun", "switch package managers", "replace npm", or "use bun instead". Includes conversion tables and lock file migration. Core conventions auto-loaded via SessionStart hook; invoke /dev-conventions:bun-tooling for full reference.

Audit project dependencies for known vulnerabilities using uv audit (Python) and bun audit (JS/TS). Use when user says "audit dependencies", "check for vulnerabilities", "security audit", "check deps", "are my packages safe", "CVE check", "dependency security", "vulnerability scan". Invoke with /dev-conventions:dep-audit.

Documentation conventions: last-updated dates, lowercase filenames, internal/ subfolder for non-shared docs, session logs, and documenting the "why". Use when creating or updating documentation, READMEs, or design docs. Invoke with /dev-conventions:doc-conventions. Triggers on "write docs", "update README", "document this", "add a design doc", "session log".

Detailed Python/uv conversion reference. Core conventions auto-loaded via SessionStart hook; invoke /dev-conventions:python-tooling for full conversion tables. Use when you need the complete uv command mapping, version pinning strategy, or lock file workflow.

Red/green TDD workflow: write a failing test first, then implement, then refactor. Use when writing tests, implementing features test-first, or when user says "TDD", "test first", "red green refactor", "write a test", "test-driven". Invoke with /dev-conventions:tdd-workflow.

Write and edit in GOV.UK / GDS house style — plain English, active voice, front-loaded content, sentence case, and no bold or italics for emphasis. Use when writing or editing reports, research write-ups, guidance, documentation, summaries, or any prose where clarity and accessibility matter.

Orchestrate end-of-session cleanup: draft a session log entry, sync bundled references, and flag any plugin-content changes that need a version bump before commit. Use when the user says "finish session", "wrap up", "end of session cleanup", "close out this session", or before committing a substantive work session. Invoke with /skill-maintainer:finish-session.

Set up persistent maintenance config and state tracking in a skills repo. Creates .skill-maintainer/ directory with config, state, and best practices checklist. Use when user says "init maintenance", "set up maintenance", "initialize skill-maintainer", "configure maintenance", "add maintenance to this repo". Invoke with /skill-maintainer:init-maintenance.

Full maintenance pass for a skills repo: pull upstream docs, sync tracked source repos, run quality checks, and review best practices for updates. Use when user says "maintain", "maintenance pass", "check upstream", "pull sources", "review best practices", "run maintenance". Invoke with /skill-maintainer:maintain.

Quick quality check for all skills in a repo: spec compliance, token budget, body size, freshness, and description quality. Use when user says "check quality", "quality report", "check my skills", "validate skills", "skill health", "are my skills ok". Invoke with /skill-maintainer:quality.

Sync the working-copy best_practices.md to the plugin-bundled reference so new `skill-maintain init` runs pull the latest rules. Use when the user says "sync bundled reference", "update bundled best_practices", "sync plugin reference", or after editing `.skill-maintainer/best_practices.md`. Invoke with /skill-maintainer:sync-bundled-ref.

Bump a plugin's version across all sources (plugin.json, marketplace.json, SKILL.md, pyproject.toml) atomically. Use when the user says "sync versions", "bump version", "align versions", or "/sync-versions <plugin> <version>". Pass plugin name and target version as arguments.

Enforces a single rule: every path in repo content must be relative to the repo root. Anything resolving outside the repo (other repos on disk, ~/.claude/<plan>, /Users/<name>/..., /home/<name>/..., $HOME-based paths) is a leak. Ships a SessionStart directive plus pre-commit and commit-msg git hooks that hard-block. Use when user says "scan for path leaks", "check for leaked paths", "are we leaking my home path", "scrub external paths", "install path-privacy hooks", "find absolute paths in this repo", "remove $HOME references", "block path leaks", "audit before commit", "privacy scan paths".

Query the agent-state DuckDB (~/.claude/agent_state.duckdb) via MCP tools instead of the `agent-state` CLI. Use when you need run history, watermarks, skill versions, or flywheel metrics. TRIGGER on thoughts like "let me run agent-state status", "I'll check the agent_state database", "show me recent runs", "agent-state runs", "agent-state tree", "agent-state watermarks", "agent-state flywheel", "what failed recently", or "list skills by domain". The MCP tools are structured, faster, and permission-managed; the CLI is for interactive debugging only.

Enable the agent-state MCP server in this repo's .mcp.json by promoting the entry from `_available_servers` into `mcpServers`. Use when user says "enable agent-state mcp", "turn on agent-state", "wire up the DuckDB MCP", "enable agent state", or "activate agent-state-mcp". Invoke with /agent-state-mcp:enable.

Design and implement Kimball-style star schemas in DuckDB for LLM agent state persistence. Use when user needs to track agent execution, model operational data, design fact/dimension tables, implement SCD Type 2, generate surrogate keys, or build analytical views. Also triggers on "star schema", "dimensional model", "DuckDB schema", "fact table", "dimension table", "SCD Type 2", "surrogate keys", "data warehouse for agents", "design a schema for", "help me track agent state", "store execution data in DuckDB", "I need to persist agent history", or "how do I model".

Browse AWM-1K catalog of 1000 pre-built tool environments by category or keyword. Use when user says "browse environments", "search AWM catalog", "find a tool environment", "list environment categories", or wants to materialize a pre-built MCP backend.

Generate database-backed MCP tool environments from scenario descriptions or browse 1000 pre-built environments. Use when user says "generate a tool environment", "create an MCP backend", "scaffold database tools", "build a sandbox", "synthesize an environment", "launch an environment", "browse environments", "AWM catalog", "create API tools", "generate test tools", "tool environment for", "database-backed tools", "sandbox for agent testing", or "create tools for".

Generate a new database-backed MCP tool environment from a scenario description. Use when user says "forge an environment", "generate tool environment", "create an MCP backend", "scaffold database tools", "build a sandbox", or "synthesize an environment from scratch".

Materialize and start a generated or catalog environment as a FastAPI+MCP server (Phase 2). Use when user says "launch environment", "start the server", "run the environment", or wants to bring up a materialized env-forge environment.

Run verification functions against current DB state for a task (Phase 2). Use when user says "verify task", "check task completion", "run verifiers", or wants to validate that an agent completed an env-forge task correctly.

Break down a goal, process, or workflow into MECE components with Agent SDK mapping. Use when user says "decompose", "break down this process", "MECE analysis", "create a decomposition tree", or pastes a JSON/YAML/CSV workflow export to analyze.

Generate Claude Agent SDK Python code scaffolding from a validated MECE decomposition. Use when user says "export to Agent SDK", "generate agent code", "create SDK scaffolding", or wants to turn a decomposition tree into runnable Python code.

Extract process knowledge from an SME through structured MECE interview. Use when user says "interview me about a workflow", "help me map this process", "extract process knowledge", or wants to build a decomposition through guided conversation.

MECE decomposition methodology, scoring rubrics, and Agent SDK mapping for process analysis. Loaded automatically when decomposing goals, tasks, processes, or workflows into Mutually Exclusive, Collectively Exhaustive components. Provides the domain knowledge used by /decompose, /interview, /validate, and /export commands. Use when user says "decompose", "break down this process", "MECE analysis", "interview me about a workflow", "map process to agents", "validate decomposition", or "export to Agent SDK".

Validate a MECE decomposition for compliance, structural integrity, and SDK readiness. Use when user says "validate decomposition", "check MECE compliance", "run validation", or wants to verify a decomposition tree passes quality gates.

Design functional, readable terminal UIs with proper visual hierarchy, semantic color, and responsive layout. Use when building CLI tools, interactive prompts, dashboards, or any terminal output with Rich, Textual, Questionary, or Click. Triggers on "terminal UI", "TUI", "Rich table", "CLI output", "terminal dashboard", "questionary prompt", "make this look better in the terminal".

Manages the read-it-later lifecycle with decision frameworks for inbox triage. Use when the user wants to process their reading inbox, batch-triage saved items, or assess inbox health metrics.

Synthesizes knowledge from your Readwise Reader library by prioritizing highlights and annotations across documents. Use when the user asks for references from their reading, wants to surface saved knowledge, or needs cross-document synthesis.

Decomposes natural language queries into targeted searches across Readwise Reader documents, highlights, and tags. Use when the user searches their reading library, asks what they saved about a topic, or wants to find specific highlights.

Show an interactive dashboard of all tracked skills, plugins, and repo hygiene in this project. Use when the user says "show skill dashboard", "skill status", "show me skill health", "what skills are stale", "check skill freshness", "skill budget overview", "show dashboard", or "run quality checks". Renders an interactive MCP App with pass/fail indicators for spec compliance, description quality, token budgets, freshness, plugin checks, and repo hygiene.

Analyze, polish, and manage Claude Code plugins. Use when user wants to evaluate a plugin (/plugin-toolkit:analyze), add standard utility commands (/plugin-toolkit:polish), or add/remove/modify plugin features (/plugin-toolkit:feature). Also use when user mentions "plugin analysis", "plugin review", "add command to plugin", or "improve plugin". Also use when user says "review my plugin", "check plugin quality", "what's wrong with my plugin", "add a help command to my plugin", or "improve my plugin structure".

Auto-generate and keep markdown documentation in sync with code using cogapp. Use when user says "keep docs in sync", "regenerate readme", "embed --help output", "run cog -r", "documentation is out of date", "sync CLI help into README", or when documentation should be derived from code rather than maintained by hand.

Guides building MCP Apps with interactive UIs, covering SDK patterns, resource registration, and host integration. Use when the user asks to "create an MCP App", "add a UI to an MCP tool", "build an interactive MCP View", or "scaffold an MCP App".

Converts OpenAI Apps SDK applications to MCP Apps SDK with step-by-step migration guidance and API mapping tables. Use when the user asks to "migrate from OpenAI Apps SDK", "convert OpenAI App to MCP", "port from window.openai", or "migrate from skybridge".

Guide JSON query tool selection and syntax. Use when user needs to extract data from JSON files, search JSON for keys, query nested JSON structures, filter JSON arrays, or process large JSON. Triggers on "query JSON", "extract from JSON", "search JSON", "parse JSON file", "jq alternative", "jsongrep", "jg command", "find in JSON", "JSON path query", "grep JSON", "large JSON file", "process JSON", "JSON extraction".