yusufkaracaburun
UserThe agile lifecycle as Claude Code + Cursor primitives. Stack-agnostic. One kit ships skills, subagents, slash commands, hooks, rules, and a Claude Code plugin manifest — for the two hosts ai-kit supports: Claude Code and Cursor.
Categories
Indexed Skills (31)
checkpoint
Save a session-continuation memo to project memory before /clear, /compact, a long break, or end of session. The next session reads the memo and resumes without context-recovery overhead. Use when context is getting big, before clearing or compacting, before pausing work, or when wrapping up for the day.
docs-sync
Content-drift checks for the current project — dead markdown links today; repo-hygiene + finished-work cleanup added in follow-up issues. Standalone counterpart to /ai:hygiene (install + catalog health). Use when wrapping up a session, after a rename or refactor, before opening a PR, or on 'check docs', 'are the links still good', 'is anything stale'. Reports only; never auto-edits files.
handoff
Compact the current conversation into a handoff document for another agent to pick up. Use when context is getting full, before /clear, or when handing work to another machine or teammate. Pass `mid-session` to compact without ending the current session.
onboard
Generate an ONBOARDING.md for a new dev or agent arriving at this project, then publish it via the harness `ShareOnboardingGuide` tool so it can be shared with a short link. Use when a teammate joins the repo, when bootstrapping a fresh dev machine, or when an agent will pick up work without prior conversation history. Re-runs idempotently — refreshes the existing guide instead of creating a new one.
rename-housekeeping
Rewrite stale absolute-path references in memory files after a local repo rename or move. Scans `~/.claude/**`, project `CLAUDE.md`/`AGENTS.md`/`.agents/memory/**`/`.planning/**`/`docs/**`/`.claude/**`, and `~/.cursor`/`~/.codex`/`~/.gemini`. Snapshots writes to `~/.claude/.backups/<ts>/` first. Use when user says "I renamed the repo", "moved this project", "verplaatst", "hernoemd", "fix stale paths", or invokes `/ai:rename-housekeeping <old> <new>`.
resume
Find the most recent /ai:checkpoint memo for the current project, summarise it, and propose the first concrete action — pick up where you left off. Use at the start of a fresh session, after /clear or /compact, when the user asks 'where was I?', wants to continue previous work, or resume after a break.
phase
Read or flip the project lifecycle phase (development | production) without re-running /ai:setup. Use when the user wants to switch a project from dev to prod (or back), check the current phase, or asks "what lifecycle is this project in", "set this to production", "we're greenfield now", or invokes `/ai:phase`.
zoom-out
Tell the agent to zoom out and give broader context or a higher-level perspective. Use when you're unfamiliar with a section of code or need to understand how it fits into the bigger picture.
autonomous
Autonomous backlog grinder — drains issues labeled `ready-for-agent` one at a time, fresh agent context per issue, persistent progress log. Wraps tdd → review → ship. Use when the user wants an AFK queue-drain run, references "Ralph pattern", or invokes `/ai:autonomous`.
recommend-rules
Recommend canonical ai-kit rules + stack-specific community rules for the current project. Use when user asks "which rules should we enable", "set up rules for this repo", or after /ai:setup to refine the default rule set. Combines deterministic scoring against detected stack with optional web search for community-curated rules.
recommend-tools
Recommend optional companion tools (graphify, caveman, llm-wiki) and stack-specific runtime helpers — MCP servers + Claude Code hooks — for the current project. ai-kit writes integration glue, never auto-installs. Use when the user asks "should I add graphify / caveman / a self-maintaining wiki", "which MCP servers fit this stack", "any hooks I should turn on", "set up companion tools", "fewer tokens", or after /ai:setup to layer extra capability.
setup
Stack-agnostic ai-kit onboarding — one setup-mode question, fast Tier-A path by default, optional full branches for Docker, tracker, domain, architecture, Sandcastle. Claude Code and Cursor. Use when the user wants to install ai-kit in a repo, set up ai-kit, onboard a project to ai-kit, re-run setup to extend Tier A → Tier B, or invokes `/ai:setup`. Re-runs idempotently — the `.ai-kit-setup` marker triggers keep/change/skip per branch instead of overwriting.
ship
Deploy a release to production or staging — pre-flight checks, release notes, deploy via project CI or docs, post-deploy smoke test. Use when landing an increment or closing a sprint increment.
triage
Triage issues through a state machine driven by triage roles. Use when user wants to create an issue, triage issues, review incoming bugs or feature requests, prepare issues for an AFK agent, or manage issue workflow.
doc-to-skill
Convert a PDF / EPUB / DOCX / RTF / ODT / MD into a SKILL.md scaffold with chapter index, claim placeholders, sources list, and provenance credit. Use when the user wants to turn a book / spec / long-form doc into a project-scoped skill, or says "make a skill from this PDF", "convert this doc to a skill", "scaffold a skill from this paper".
contribute-eval
Turn a skill failure into a regression test by contributing a new eval case (prompt fixture + golden) to the ai-kit eval suite. Use when user says "this skill failed", "ai-kit got this wrong", "add a test for this", "contribute eval", "/ai:contribute-eval", or whenever an ai-kit skill output is materially worse than the user expected and the user wants the fix to stick.
diagnose
Disciplined diagnosis loop for hard bugs and performance regressions. Reproduce → minimise → hypothesise → instrument → fix → regression-test. Use when user says "diagnose this" / "debug this", reports a bug, says something is broken/throwing/failing, or describes a performance regression.
feedback
Send structured feedback about ai-kit itself — friction, surprise, clarity, gaps, or wins — to the ai-kit GitHub repo. Use when user says "this skill was annoying", "ai-kit should also do X", "report this to ai-kit", "give feedback on ai-kit", "/ai:feedback", or any variant of "this kit could be better at …".
grill-me
Interview the user relentlessly about a plan or design until reaching shared understanding, resolving each branch of the decision tree. Use when user wants to stress-test a plan, get grilled on their design, or mentions "grill me".
grill-with-docs
Grilling session that challenges your plan against the existing domain model, sharpens terminology, and updates documentation (CONTEXT.md, ADRs) inline as decisions crystallise. Use when user wants to stress-test a plan against their project's language and documented decisions.
improve-codebase-architecture
Find deepening opportunities in a codebase, informed by the domain language in CONTEXT.md and the decisions in docs/adr/. Use when the user wants to improve architecture, find refactoring opportunities, consolidate tightly-coupled modules, or make a codebase more testable and AI-navigable.
next
Ranked "what's next" list for the current project — combines open GitHub issues (P0-P3 + status:in-progress + ready-for-agent labels), roadmap-only tasks, and setup-deferred branches into one prioritized backlog. Use when the user asks "wat is open", "what's next", "wat moet ik nu doen", "openstaande taken", "status van het project", "hoogste prioriteit", or invokes /ai:next.
prototype
Build a throwaway prototype to flesh out a design before committing to it. Routes between two branches — a runnable terminal app for state/business-logic questions, or several radically different UI variations toggleable from one route. Use when the user wants to prototype, sanity-check a data model or state machine, mock up a UI, explore design options, or says "prototype this", "let me play with it", "try a few designs".
qa
QA a web app at a given URL — smoke test critical paths, report bugs with repro steps, optionally fix and re-verify. Use when validating a running build, after a deploy, or before handing off to stakeholders. Prefers Playwright or browser MCP when available.
retro
Sprint or cycle retrospective — what shipped, what blocked, stakeholder feedback, learnings, next backlog. Use at end of a sprint or release cycle. Works for Scrum and Kanban.
review
Pre-merge code review on the current branch diff. Checks correctness, security, test coverage, and alignment with CONTEXT.md. Use before landing, opening a PR, or pre-release security checks.
setup-gh-workflow
Install ai-kit's GitHub workflow hygiene — issue templates with DoR/DoD checklists, two workflows (close→DoD check, label→DoR check, auto-promote Todo→Ready), and P0-P3 + epic/* + area/* + status:in-progress labels. Tier-A — fires from `/ai:setup` when a GitHub remote is detected. Dutch default; `--lang en` for English. Use when user wants to set up GH issues + Projects with enforced DoR/DoD, or invokes `/ai:setup-gh-workflow`.
should-i-use
Assess whether a tool, library, repo, URL, article, pasted text, or image is worth adopting in the current project. Produces a structured verdict — vendor, wire, adopt-as-pattern, or ignore — with status check, value, cohesion impact, and a recommendation. Use when the user shares a tool / repo / link / screenshot and asks "should I use this", "beoordeel dit", "wat vind je hiervan", "is dit iets voor ons", or whether something belongs in the project.
tdd
Test-driven development with red-green-refactor loop. Use when user wants to build features or fix bugs using TDD, mentions "red-green-refactor", wants integration tests, or asks for test-first development.
to-issues
Break a plan, spec, or PRD into independently-grabbable issues on the project issue tracker using tracer-bullet vertical slices. Use when user wants to convert a plan into issues, create implementation tickets, or break down work into issues.
to-prd
Turn the current conversation context into a PRD and publish it to the project issue tracker. Also supports Defer mode — when the user wants a short parking-lot capture ("defer this", "track this for later", "follow-up on …", "we'll do this in N") instead of a full PRD, produce a one-line issue + roadmap entry. Use when user wants to create a PRD, defer a lead, or capture a follow-up.
Bio shown is the top-scored skill's repo description as a fallback — real GitHub bios land in a future update.