alexngai

Find the commit that introduced a regression automatically with git bisect run.

Configure Jest for a TypeScript project — ts-jest, module mapping, and coverage thresholds.

Diagnose a slow Postgres query and add the right index using EXPLAIN ANALYZE.

Test React components with Testing Library — render, query by role, and assert on user-visible behavior.

Fix TypeScript ES module import errors by adding .js extensions to relative imports.

Migrate a Jest test suite to Vitest — config, globals, and mock API differences.

Comprehensive guide for creating high-quality skills for Claude. Use when building a new skill, writing a SKILL.md, designing skill frontmatter, planning skill structure, debugging trigger issues, or reviewing an existing skill for improvements. Covers standalone skills, MCP-enhanced skills, and skill-tree programmatic skills.

Helps write clear and conventional git commit messages.

Launch a Codex agent team from an openteams YAML topology. Accepts a topology name (e.g. "gsd", "bmad-method") or path to a custom YAML topology, generates role artifacts, and coordinates the team through opentasks (shared task graph) and agent-inbox (inter-agent messaging).

Inspect the coordination MCP surface swarm-codex registers for role threads. Reports which daemons (opentasks, agent-inbox, minimem) are available, which are actually running, and what's wired into the per-run CODEX_HOME `config.toml`. Non-invasive — never modifies the user's `~/.codex/config.toml`.

Inspect and optionally install MCP providers declared by the current swarm team. Non-invasive by default — reports status without modifying user MCP configuration unless explicitly requested.

Launch an agent team from an openteams topology template. Creates a native Claude Code team and spawns all agents directly.

Use when managing work items, linking external data (Slack, docs, URLs) to tasks, querying task dependencies or blockers, leaving cross-system feedback on specs or issues, or coordinating work across multiple task systems. Use when the agent has access to the `opentasks` CLI.

Search and manage memories using minimem. Use when the user wants to recall previous context, decisions, notes, or stored information. Also use when storing new information for later retrieval.