website

Generate and maintain module-level README documentation for any application, with automatic language detection (Go, TypeScript, Python, Rust, Java) and CLAUDE.md documentation index management. Use when users want to document their codebase, generate READMEs for modules/packages, create a documentation index, update existing docs to match current code, or run "document this project". Triggers on requests like "document the app", "generate docs", "update documentation", "create doc index", "add docs to CLAUDE.md".

Skill for writing rigorous financial analysis articles about publicly listed companies (equity research, deep dives, company teardowns). Use when the user asks to write or draft a company analysis article, stock analysis, equity research note, or company deep dive — in French or English. Supports three output modes selectable at launch: X thread, Substack newsletter, or Hartza Capital blog post. Triggers on requests like 'écris une analyse de [société]', 'write a company analysis for [ticker]', 'analyse de société cotée', 'deep dive [company]', 'article sur [stock]', 'analyse financière [entreprise]'.

Automatically detect and remove dead code from any project by analyzing the codebase and dispatching to the appropriate language expert agent. Use when user asks to "clean dead code", "remove unused code", "find dead code", "nettoie le code mort", "supprime le code inutilisé", or "cleanup unused". Supports Go, Python, Rust, TypeScript/React, and Java projects.

Work on an existing GitHub issue by its number. Use when user asks to "work on issue #123", "implement issue 45", "fix issue #67", "start working on #12", "travailler sur l'issue #8", "implementer le ticket #5", or "prendre en charge l'issue". Reads the issue, analyzes context, creates a feature branch, dispatches to the appropriate specialized agent, and prepares a pull request. Complements the github-issue skill (which creates issues).

Automatically analyzes multi-language projects to identify improvement opportunities across functional, performance, security, quality, architecture, and testing dimensions. Use when user asks to "analyze the project for improvements", "what can be improved in this codebase", "audit the project", "review code for optimization", "find improvements", "analyser le projet", "améliorer le code", or "audit du code". Detects Go, Python, Rust, TypeScript/React, and Java, launches appropriate expert agents in parallel, and generates an actionable implementation plan.

Automatically generates comprehensive test suites for multi-language projects by parallelizing test creation across Go, Python, Java, Rust, and TypeScript/React. Use when user asks to "write tests", "add tests", "generate tests", "test coverage", "create test suite", "écrire des tests", "ajouter des tests", or "générer des tests". Detects all languages in the project, spawns parallel test generation tasks via qa-test-engineer agent, and validates functionality with full test execution.

Orchestrate the full release process for any project using Makefile targets — changelog generation (via the changelog skill), version bump, git tagging, build, artifact upload, and GitHub Release creation. Use when the user says "release", "publish version", "cut a release", "ship it", "make a release", or wants to prepare and publish a new stable or beta version. Requires a Makefile at project root. Argument is the target version (e.g., "1.2.3" or "1.2.3-beta.1").