This repository also configures its own agents. See what agent-skills tells them →
Skill Claude CodeCodex
Save coding-agent Q&A conversations (Claude Code, Codex, omp, ...) as structured Obsidian-compatible knowledge notes, or rewrite fact-checks into corporate-friendly Teams publish format. Use when user wants to persist conversation content as a markdown file, or rewrite analysis for team sharing.
Skill Claude Code
Use when asked to sync, pull latest, update a repo, or show what changed recently. Keeps any git repo up-to-date by pulling the latest changes and optionally syncing submodules. Also generates a "What's New" digest of recent changes. Works in a single repo OR from a non-git parent directory to batch-sync all…
Skill Claude CodeCodex
Use when generating status reports, project summaries, retrospectives, or any structured report from work logs, git history, and activity records. Supports weekly, monthly, project, sprint, incident, and retrospective report types.
Skill Claude Code
Cross-validate technical claims using multiple independent verification strategies (official docs, counter-examples, source code inspection). Use when a tech-feasibility report, design document, or any technical analysis contains claims that were verified by only one method or one AI session. Triggered by…
Skill Claude CodeCodex
Use after running 2+ research skills (critical-research, tech-feasibility, narrative-auditor, codebase-audit) to synthesize findings into a unified decision document. Resolves conflicts between sources, weighs evidence, and produces an actionable recommendation.
Skill Claude CodeCodex
Use when repeated code review, doc review, design spec review, implementation plan review, or multi-agent review loops need severity gates, stop conditions, and implementation-readiness verdicts to prevent over-review, reviewer drift, duplicate findings, minor/nit churn, or infinite Claude/Codex/Gemini cycles.
Skill Claude CodeCodex
PM (Product Manager) role for the multi-agent role system. Produces structured requirements artifacts calibrated by project size (small/medium/large). Invoked by role-orchestrator via subagent dispatch, or directly for standalone requirements work. Triggered by "PM analysis", "requirements gathering", "role-pm", or…
Skill Claude CodeCodex
RD (Developer) role for the multi-agent role system. Produces structured design artifacts calibrated by project size (small/medium/large), taking PM requirements as input. Invoked by role-orchestrator via subagent dispatch, or directly for standalone design work. Triggered by "RD analysis", "technical design"…
Skill Claude CodeCodex
Use when the user explicitly invokes $session-preferences or /session-preferences to enable a modular policy bundle for the current session, or to propose and confirm a new rule. Never enable automatically.
Skill Claude CodeCodex
Use when reviewing SKILL.md files, checking skill quality standards, or deciding whether to share a skill with a team.
Skill Codex
A skill for finding and choosing other available skills. It can also recommend workflows when a request could match more than one skill.
Skill Claude CodeCodex
A synchronisation tool for copying Claude skills to other coding assistants such as Codex, Gemini, Cursor, or Antigravity.
Skill Claude CodeCodex
Analyze PM spec and UI wireframe from an RD perspective to find gaps, ambiguities, and undefined edge cases before implementation starts. Produces a structured pre-dev question list ready for a single alignment meeting with PM/Designer. Use when: (1) RD receives a new feature spec or wireframe and wants to identify…
Skill Claude CodeCodex
Use when evaluating whether a specific technology, tool, or approach can solve a given problem before committing to implementation. Triggered by questions like "can X do Y?", "will X work for Y?", "is X feasible for Y?", or "should I use X to achieve Y?".
Skill Claude CodeCodex
Orchestrates a full technical research workflow by chaining specialized skills in sequence: brainstorming → tech-feasibility → assumption-extractor → micro-poc-validator → critical-research → narrative-auditor → research-cross-validator → research-synthesis. Use when evaluating a technology choice, planning a…
Skill Claude CodeCodex
Analyze UI/UX design from screenshots. Use this Skill when the user asks to analyze UI design, review interface screenshots, check design quality, evaluate user experience, or get design improvement suggestions.
Skill Codex
Use when the user wants to query their work journal, track task progress, or understand how decisions evolved over time. Analyzes activity-logger records and work logs to answer questions about project history, decisions, TODOs, and timeline evolution.
Skill Claude CodeCodex
Use when recording work activities, preserving explicitly consulted document paths, or preparing activity data for cross-session aggregation.
Skill Claude CodeCodex
Orchestrates a multi-role development pipeline (PM → RD) using subagent dispatch. Reads project-profile.yaml to calibrate role behavior by project size. Each role runs in an isolated subagent to prevent context pollution. Includes gate checks for user approval between phases. Triggered by "role orchestrator"…
Skill Codex
Use when asked to humanize, de-AI, unslop, or strip AI flavor from any text, or when writing or revising fiction, release notes, announcements, PR/issue replies, code-review comments, postmortems, tickets, technical articles, or blog posts that must not read as machine-written. Make AI-generated writing read as…
Skill Claude CodeCodex
Use when the user asks to review a technical wiki, design proposal, RFC, or enhancement proposal while lacking project context ("review this enhancement proposal", "I don't know this project", "help me understand this technical wiki", "review this RFC", "technical proposal review"), or wants informed feedback on a…
Skill Claude CodeCodex
Use when AI-assisted work may involve secrets, customer data, sensitive internal information, risky AI-generated output, or repository commits — inspect prompts, files, screenshots, logs, source code, and staged changes rather than trusting repository classification; sanitize sensitive data before AI processing…
At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: