huuanh20/awesome-ai-agent-skills
Agent Claude Code
Plan-creation sub-agent for the /plan pipeline. Given a feature description and optional research reports, writes the full plan directory: plan.md overview + one phase-XX-{name}.md per phase. Used in Fast, Hard, Parallel, and Two modes.
huuanh20/awesome-ai-agent-skills
Agent Claude Code
General-purpose Playwright screenshot agent. Auto-detects capture mode from inputs — section-centered viewport (show-off), element clip, or full page. Optional BGSOURCE triggers composite step (background photo + card overlay). Spawned by /show-off and any task needing screenshots.
huuanh20/awesome-ai-agent-skills
Agent Claude Code
Finalize sub-agent used by /cook and /fix. Syncs completed phase checkboxes in plan.md, updates plan Status, and reports overall progress. Skipped gracefully if no plan.md exists.
huuanh20/awesome-ai-agent-skills
Agent Claude Code
Independent senior-engineering quality auditor for the ck:quality skill. Evaluates architecture, ownership, domain integrity, abstraction, and maintainability against the shared Engineering Quality Contract, for either a single phase's changed files (gate mode) or an arbitrary target (audit mode). Never edits files…
huuanh20/awesome-ai-agent-skills
Agent Claude Code
General-purpose research sub-agent. Given a topic and an assigned role, investigates one angle and returns a structured report. Spawned in parallel pairs (Primary + Alternative) by any pipeline needing option comparison. Budget ≤5 tool calls.
huuanh20/awesome-ai-agent-skills
Agent Claude Code
Evidence-gathering agent for the /fix pipeline. Given a bug description, greps for error patterns, reads affected source files, and checks recent git changes. Returns a structured evidence report within a ≤6 tool call budget.
huuanh20/awesome-ai-agent-skills
Agent Claude Code
Testing sub-agent for the ck:test skill. Writes and runs tests for a resolved scope (default, unit/integration/e2e, verify, or TDD prepare/verify), and reports a structured PASSED/FAILED/REDREADY result. Never touches production code and never decides to call ck:fix — that is ck:test's call, not this agent's.
huuanh20/awesome-ai-agent-skills
Command Claude Code
Explore and debate solutions before writing code — scout codebase, ask clarifying questions, research options, and produce a structured decision report. Ends with → /ck:plan (implement) or /ck:journal (archive). No code is written.
huuanh20/awesome-ai-agent-skills
Command Claude Code
Code review — local uncommitted changes or GitHub PR (pass PR number/URL for PR mode).
huuanh20/awesome-ai-agent-skills
Command Claude Code
Set Claude's coding explanation level for this session/project (-1=Default → 5=God Mode).
huuanh20/awesome-ai-agent-skills
Command Claude Code
Implement a plan phase by phase. Cook never writes or runs tests; every phase requires an APPROVED ck:quality gate before completion. Modes: --fast or --hard. Flag: --tdd. --json selects the phased master plan.json; --plan selects a Markdown plan.md.
huuanh20/awesome-ai-agent-skills
Command Claude Code
Generate a concise FE handoff doc for changed endpoints — branch, APIs, request/response contracts, search params, error codes.