mohammadhprp/teammate/bootstrap-project
Claude Code and agent skill by mohammadhprp/teammate
- community
Checking that the SKILL.md file is still in the repository.
Prepare a target project before delegating: use `find-skills` to install the skills the work needs, then create or update the project's `AGENTS.md` (and `CONTEXT.md`) with its stack, commands, conventions, and acceptance bar. Use before the first delegation into a project, when a new project has no `AGENTS.md`, when the work needs a capability the project lacks (for example web design, testing, or a framework), or when a worker's output was thin because the project's context was. Do this first, not after a bad result.
Install
npx vskill@latest install mohammadhprp/teammate/bootstrap-projectSKILL.md
The same SKILL.md was published earlier as mohammadhprp/teammate/bootstrap-project.
Bootstrap a project
A worker can only be as good as the project it lands in. A fresh project has no AGENTS.md, no conventions, and none of the skills the work needs — so the worker guesses, the review has nothing to hold it to, and the result is generic. Fix that before delegating: give the project its context and its skills.
When to use
- Before the first delegation into a project, or any time a project has no
AGENTS.md.
- When the work needs a capability the project does not have — web design,
framework specifics, testing, deployment, and so on.
- When a worker's output was weak and the cause was a thin project context,
rather than the worker.
When not to use
- The project already has a solid
AGENTS.mdand the skills this work needs. - A trivial, well-specified change in an established project.
Inputs
- The resolved project root (
multi-project-context). - The work to be done, so you know which domain and stack matter.
- Any existing
AGENTS.md,CONTEXT.md, and project skills
(load-project-context).
Procedure
- Resolve and read. Pin the project root, then read what already exists:
AGENTS.md, CONTEXT.md, local skills, and the commands. Never overwrite the developer's context; merge into it. Record a durable allow rule for the project boundary (OpenCode applies it from the next session, so do this before the primary needs to read or write the project):
``bash python3 scripts/tm.py permissions allow --cwd "<project-root>" ``
- Set up the harness surface. Confirm the resolved harness and render its
adapter files into the project, so the worker's harness can load the agent definitions and skills:
``bash python3 scripts/tm.py harness python3 scripts/tm.py skills sync --cwd "<project-root>" python3 scripts/tm.py agents sync --cwd "<project-root>" ``
- Detect the work's domain and stack. From the goal and the repository
(manifests, file types, existing config), name the stack and the capabilities the work needs. "A landing page" needs visual design and front-end skill; "a CLI" needs testing and packaging — be specific about the query.
- Find skills with the Skills CLI. Use the
find-skillsskill when it is
available — it is the guide to npx skills — and check the skills.sh leaderboard first. Then search:
``bash npx skills find "<domain> <task>" ``
Verify before recommending: install count (prefer 1K+, be wary under 100), source reputation, and the source repository. Do not pick on keywords alone.
- Install what the work needs, into the project. Install project-scoped so
the capability travels with the project and does not leak between projects; use global (-g) only for something broadly useful everywhere:
``bash cd "<project-root>" npx skills add <owner/repo@skill> -y ``
Installing new skills is additive; if a candidate would replace an existing project skill, stop and ask instead. Remember the project's harness skills directory (from tm harness; for example .claude/skills, .opencode/skills, or .agents/skills for codex) holds two populations — Team Mate-managed skills (tracked in .teammate-managed.json) and Skills-CLI-installed ones (tracked in skills-lock.json). A Team Mate sync never overwrites a skill it does not own, and a name collision between the two is skills-lock.json's owner to resolve, not something to silently replace. If nothing suitable exists, say so and proceed with the common and worker skills — do not install noise.
- Write or update `AGENTS.md`. Fill
templates/project-AGENTS.mdwith what
you learned: purpose, stack, the real commands (install, build, test, lint, run), the conventions, and the acceptance bar — including how a user-facing result is judged. Under Checks, name at least one validator that runs offline for this work. For a UI that means an HTML check and an accessibility check (npx html-validate, npx @axe-core/cli, or the stack's equivalent), not only a screenshot; a validator the review cannot run is not a check. For a UI, also name the dev/preview port the app is served on and require a per-worker agent-browser session (--session <stream>), so a later parallel run has a convention to follow instead of fighting over one shared browser tab. Keep any developer-authored sections; add what is missing.
- Create `CONTEXT.md` if missing. A short file naming the project's
architecture, key files, and how to work in it. Leave an existing one alone.
- Report. State the skills installed, the files written or updated, and
anything you could not determine — so the plan that follows is grounded.
Output
A project that can accept work: its AGENTS.md and CONTEXT.md in place, the harness's skills and agent definitions rendered into it, and the skills the job needs installed. Brief the worker with it and it will not have to guess.
Failure and escalation
- No suitable skill found: say so and continue with the common and worker
skills; a missing skill is not a reason to stall.
- The project already has context that conflicts with what you would write:
propose the change and let the developer decide; do not clobber it.
- A candidate skill would overwrite a project-owned one, or looks low-quality:
do not install it; list it and ask.
Security
- Security scanPassed, 90 of 100Pattern checks on the skill's files.
- Model reviewPassed
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingsrc/skills/bootstrap-project/SKILL.md
Popularity
- 7-day trend score
- 0
Not refreshed yet.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/mohammadhprp/teammate/bootstrap-project)