verygoodpluginsverygoodplugins/browser-hand/skills/vgp-repo-bootstrap/SKILL.md

vgp-repo-bootstrap

Skill2 starsChanged 4 months agoMITView on GitHub

Bootstrap or repair the standard Very Good Plugins agent-facing project surface: AGENTS.md as canonical, CLAUDE.md as @AGENTS.md, MIGRATING.md when renaming, GitHub hygiene (SECURITY, PR/issue templates, CODEOWNERS, dependabot), and product vs MCP naming discipline. Use when spinning up a new VGP repo, forking an upstream project, renaming product/CLI/skill identifiers, or when a repo still treats CLAUDE.md as the only agent contract. Trigger phrases: new project scaffold, agents.md bootstrap, claude.md to agents.md, repo hygiene, SECURITY.md, migrating file, standardize this repo, spin up a product repo.

In this repository4 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- browser-handSkill
- dev-browserSkill

The file

We split this skill into the rules it gives an agent

4 rules. About 0 are common, 1 is distinctive, and 3 are very specific to this project.

≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
3
Very specific
Tied to their project's names, files or tools. Probably skip.

Common
Distinctive
Very specific

Very specific:
Rename discipline: update agent-facing strings (skill name, CLI, README product name).
line 80
Distinctive:
No second conflicting “project rules” file that agents will load instead
line 85
Very specific:
Commit with conventional message, e.g.
line 97
Very specific:
If the repo ships a skill, admit/update it via AutoVault rather than only copying into ~/.claude/skills
line 98

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.

Files

SKILL.md

VGP repo bootstrap

Spin up (or repair) the agent + hygiene surface every VGP product or tool repo should share — without reinventing it from AutoVault, mcp-ecosystem, or memory.

This skill is not the MCP server scaffolder. For MCP servers, use mcp-ecosystem (create-server.sh / apply-templates.sh / STANDARDS.md). For everything else (CLI tools, extensions, agent skills monorepos, product forks), use this checklist.

When to load

New public or private VGP repo that agents will edit
Fork of an upstream project that still says only CLAUDE.md / old product name
Repo still has full agent rules in CLAUDE.md and no real AGENTS.md
Missing SECURITY.md, PR template, CODEOWNERS, or dependabot
Product rename (e.g. dev-browser → browser-hand) needs a MIGRATING.md

Canonical agent surface (locked)

File	Role
`AGENTS.md`	Single source of truth for coding agents (all hosts)
`CLAUDE.md`	Exactly one line: `@AGENTS.md` (Claude Code import; no duplicate body)
`MIGRATING.md`	Required when renaming product/CLI/skill or forking with a new brand
`README.md`	Humans; link AGENTS + MIGRATING when relevant
`CONTRIBUTING.md`	Humans; point agents at AGENTS.md
`.github/SECURITY.md`	Private vulnerability reporting
`.github/PULL_REQUEST_TEMPLATE.md`	Conventional-commit-friendly PR body
*`.github/ISSUE_TEMPLATE/`**	Bug + feature + security contact link
`.github/CODEOWNERS`	Default reviewers
`.github/dependabot.yml`	Dependency PRs

Reference implementations:

MCP (fleet pattern): mcp-wp, mcp-automem — CLAUDE.md is @AGENTS.md
Product (this monorepo’s parent): browser-hand after bootstrap
Deep product AGENTS body: autovault (architecture + commands; may still have a fat CLAUDE.md — prefer the import pattern for new work)

Workflow

Detect existing surface: list root AGENTS.md, CLAUDE.md, MIGRATING.md, .github/*.
Classify:
- mcp → hand off to mcp-ecosystem templates; still enforce @AGENTS.md import.
- product / fork → apply this skill’s templates.
Fill placeholders from references/checklist.md (name, slug, description, stack, commands).
Write files from templates/ (do not leave a fat CLAUDE.md body).
Rename discipline: update agent-facing strings (skill name, CLI, README product name). Leave intentional upstream package/binary names only when documented in MIGRATING.md.
Verify:
- CLAUDE.md is exactly @AGENTS.md (plus optional trailing newline)
- AGENTS.md states that it is canonical
- Security template points at the real github.com/<org>/<repo>/security/advisories/new
- No second conflicting “project rules” file that agents will load instead

Anti-patterns

Symlinking CLAUDE.md → AGENTS.md with a body titled # CLAUDE.md (breaks the import convention and confuses fleet sweeps)
Duplicating the full AGENTS body into CLAUDE.md “for Claude”
Storing secrets or machine-local paths in AGENTS.md
Treating AutoHub (or any hub) as required runtime when the product is self-contained — document hubs as optional adapters (see browser-hand docs/ecosystem-links.md pattern)
Renaming without MIGRATING.md while old skill names remain in the wild

After bootstrap

Commit with conventional message, e.g. docs: AGENTS.md canonical + GitHub hygiene
If the repo ships a skill, admit/update it via AutoVault rather than only copying into ~/.claude/skills
For MCP servers, still run mcp-ecosystem audit/apply for workflows and server.json

Placeholders used in templates

Token	Example
`{Product Name}`	Browser Hand
`{repo}`	browser-hand
`{org}`	verygoodplugins
`{cli}`	browser-hand
`{slug}`	browser-hand (AutoMem bare tag)
`{one-liner}`	Agents control real logged-in Chrome via a local extension bridge
`{old-name}`	dev-browser

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

It worked
Partly
It didn't

What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free