core

Audit TSX files for WCAG 2.1 AA accessibility compliance.

Git commit all changes and update docs

Create a changelog by comparing git refs.

Optimize documentation for conciseness and clarity.

Show status of all active plans

Use this agent when you need to create or update CLAUDE.md files for projects or specific subfolders. This agent should be used when: 1) Starting a new project that needs development guidelines, 2) Adding Claude-specific instructions to an existing codebase, 3) Creating subfolder-specific CLAUDE.md files for microservices or feature modules, 4) Updating existing CLAUDE.md files to reflect new patterns or requirements, 5) Standardizing development practices across team projects.

Deeply analyzes existing codebase features by tracing execution paths, mapping architecture layers, understanding patterns and abstractions, and documenting dependencies to inform new development

Reviews code for bugs, logic errors, security vulnerabilities, code quality issues, and adherence to project conventions, using confidence-based filtering to report only high-priority issues that truly matter

Use this agent when you need concise, context-rich technical documentation (300-500 lines) that provides architectural insights and navigation guidance. The agent focuses on documenting WHY and HOW things connect rather than duplicating code details, since LLMs have direct access to the code.

Use this agent when you need to document a feature implementation from the current session into a reusable playbook. This agent should be called after significant progress has been made on a feature and you want to capture the implementation details, context, and references for future sessions.

Audit TSX files for WCAG 2.1 AA accessibility compliance. Reports findings by severity with WCAG references and code fixes. Use when the user says 'a11y audit', 'accessibility check', 'check accessibility', 'WCAG audit', 'is this accessible', 'audit for screen readers', or asks about accessibility of changed or specific files.

Generate a changelog from git history. Compares tags, branches, or commits to produce a categorized, human-readable changelog. Handles multiple scenarios: what's unreleased on HEAD vs latest tag, tag-to-tag comparisons, staging/RC vs production, or any custom git range. Invoke with /gen-changelog or trigger naturally with phrases like 'what changed since v2.0', 'release notes', 'compare staging to prod', 'what's unreleased'.

Self-improving loop for the project's institutional knowledge. This project maintains a structured knowledge system — a root CLAUDE.md that indexes standards docs (error handling, database, security, accessibility, etc.), directory-level CLAUDE.md files for each module, and dedicated docs for cross-cutting concerns. These steering docs are what enable high-quality, consistent, secure code across every session. Invoke /learn after sessions where mistakes were made, corrections given, or non-obvious patterns validated. It scans the conversation, extracts generalizable lessons, and updates the right doc so the same mistake never happens twice. User-invoked only — never triggers automatically.

Optimize documentation for conciseness and clarity. Strengthens vague instructions with explicit criteria, removes redundancy while preserving meaning, and tightens prose. Use when the user says 'optimize this doc', 'tighten this', 'make this more concise', 'clean up this documentation', 'this doc is too long', 'remove redundancy', or asks to improve a markdown/documentation file.

Automates browser interactions for web testing, form filling, screenshots, and data extraction. Use when the user needs to navigate websites, interact with web pages, fill forms, take screenshots, test web applications, or extract information from web pages.

1 hook across 1 event