Replace ad-hoc AI prompts with a repeatable spec-driven workflow: from idea to verified spec.
Named after the tracer-bullet technique from The Pragmatic Programmer: Tracer + Kit.
Markdown skills, zero runtime deps. No build step, no project dependencies.
Why TracerKit?
AI assistants work best with small, well-scoped tasks — not sprawling layers or flat task lists. TracerKit structures every feature as tracer-bullet vertical slices: each phase cuts through every layer (schema → service → API → UI → tests) and is demoable on its own. Integration problems surface early, not at the end.
Four skills drive the workflow: define (/tk:prd), plan (/tk:plan), build (/tk:build), verify (/tk:check). The AI reads your specs directly, counts progress, and marks completed work done. Pure Markdown, zero runtime deps.
DEFINE PLAN BUILD VERIFY
/tk:prd ───▶ /tk:plan ───▶ /tk:build ───▶ /tk:check
│ │ │ │
▼ ▼ ▼ ▼
PRD doc Phased plan Implement phase Pass/fail
+ feedback loops
Get Started
Claude Code (recommended)
Marketplace install:
/plugin marketplace add helderberto/tracerkit
/plugin install tk@tracerkit-marketplace
CLI install:
npm install -g tracerkit
tracerkit init
Skills are installed to ~/.claude/skills/, available in every project. Safe to re-run.
Per-project install (team members get skills via git)
tracerkit init . # install to .claude/skills/ in current dir
tracerkit update . # update project-scoped skills
tracerkit uninstall . # remove project-scoped skills
Cursor
Copy skills into .cursor/rules/, or use Notepads for on-demand context. See docs/cursor-setup.md.
Gemini CLI
Install as native skills for auto-discovery:
gemini skills install https://github.com/helderberto/tracerkit.git --path skills
Or add to GEMINI.md for persistent context. See docs/gemini-cli-setup.md.
GitHub Copilot
Copy skills to .github/skills/ or add workflow rules to .github/copilot-instructions.md. See docs/copilot-setup.md.
OpenCode
Uses agent-driven skill execution via AGENTS.md. See docs/opencode-setup.md.
Other agents
Skills are plain Markdown with YAML frontmatter — they work with any agent that accepts instruction files. Clone the repo and point your agent at the skills/ directory.
Workflow
You: /tk:prd add dark mode support
AI: Written .tracerkit/prds/dark-mode-support.md
Run `/tk:plan dark-mode-support` next?
You: /tk:plan dark-mode-support
AI: Phase 1 — Theme visible end-to-end
Phase 2 — User can toggle and persist preference
Written .tracerkit/plans/dark-mode-support.md
Run `/tk:check dark-mode-support` when ready?
You: /tk:build dark-mode-support
AI: Phase 1 — Theme visible end-to-end (3 remaining)
Implementing... all checks pass. Commit?
You: /tk:check dark-mode-support
AI: Total: 7/7
All phases complete — implementation verified.
Between sessions, /tk:brief shows active features and picks up where you left off:
You: /tk:brief
AI: dark-mode-support (3/7)
Phase 1 — Theme visible end-to-end: 3/3
> Phase 2 — User can toggle and persist preference: 0/4
**Focus → dark-mode-support**
See Examples for full walkthroughs.
Skills
