Skip to content
Back to skills

Project Docs Bootstrap

ASecurity

Use to create or regenerate a repository''s .issueops operating documents and AGENTS.md routing block from repository evidence. Incremental refresh: project-docs-update; restructuring: project-docs-optimize.

  • 6 stars
  • 0 votes
  • 0 copies
  • 0 views
  • Added November 7, 2026
ai-agentspythongobashtestingapifrontenddocumentation

Works with

  • cli
  • api
  • mcp

Security analysis

A100/100

Pro scans all 4 files and shows the line behind each finding

Scanned November 7, 2026

npx -y skills add m16khb-org/issueops --skill project-docs-bootstrap --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Project Docs Bootstrap?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Project Docs Bootstrap
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/m16khb-org-project-docs-bootstrap/badge)](https://www.skillsdirectory.com/skills/m16khb-org-project-docs-bootstrap)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: project-docs-bootstrap
description: 'Use to create or regenerate a repository''s .issueops operating documents and AGENTS.md routing block from repository evidence. Incremental refresh: project-docs-update; restructuring: project-docs-optimize.'
---

# Project Docs Bootstrap

## Goal

Create or refresh the repo-local `.issueops` document family and the
`AGENTS.md` routing block from repository evidence. Two passes run in order:
the deterministic static pass (CLI engine), then the agent enrichment pass
(`PROMPT.md` in this skill). This skill is the docs dimension of the
`project-bootstrap` orchestration and is also usable standalone.

## Use and Boundaries

issueops is an open-source library applied to many different projects.
Defaults and templates are starting points, never mandates: a target repo's
own conventions, doc language, and curated content are always authoritative.

Use for first-time creation of `.issueops` docs, or an explicit template
refresh of existing generated docs. Route instead:

- incremental refresh while work is happening → `project-docs-update`;
- oversized or badly structured docs → `project-docs-optimize`;
- a repo whose `.issueops` family roots are not modular yet (no
  `documentation/manifest.json`) → `project-docs-bootstrap` creates the modular
  layout; there is no flat-layout path.

If `.issueops` documents already exist and the user did not ask for a
refresh, do not run the write pass; route to `project-docs-update`.

## Inputs

- repository root (defaults to the current directory);
- `--dry-run` for a plan without writing;
- `--sync` to replace existing generated docs from current templates and
  evidence — only on explicit request.

## Safety Rules

1. Never overwrite an existing `AGENTS.md` wholesale. The engine may prepend
   the behavioral block when missing and manages only the `ISSUEOPS`
   marker block after that.
2. Run the dry-run plan first when any doubt exists about what will change.
3. Without `--sync`, existing generated docs are preserved; report the
   `sync_available` warning instead of forcing a refresh.
4. Treat static output as an evidence-backed draft. If the project already
   has stronger local docs, preserve and reference them.
5. The enrichment pass must not invent facts: anything not confirmed by
   source files, command output, or explicit user instruction is marked
   `Unknown / not confirmed` with a way to verify.
6. Respect the repo's established conventions and language. Existing
   `AGENTS.md` behavioral guidance, doc style, and terminology win over
   harness defaults unless the user explicitly overrides them; enrich in the
   dominant language of the existing docs. Never stack generic template text
   over repo-authored rules.

## Static vs Agent-Filled Boundary

The static pass fills only: document skeletons, detected
languages/package managers, candidate test/build/lint commands, routing,
generic safety/testing/API guidance, `AGENTS.md` managed blocks, and repo
profile metadata.

The agent enrichment pass fills from codebase evidence: actual architecture
(entrypoints, boundaries, data flow, auth/error surfaces), actual operations
(setup, env, build/test/lint/deploy, CI), actual conventions and testing
guidance, and concrete cautions/ADRs. Standard engineering topics (DDD, clean
code, layered/hexagonal architecture, SOLID, OOP, testing, Swagger, exception
handling, ...) are evaluated against that evidence using
`references/engineering-standards.md`; only confirmed topics reach the docs.
See `PROMPT.md` for the full fill targets per document.

## Workflow

1. Plan without writing:

   ```bash
   issueops project bootstrap --repo . --dry-run --json
   ```

   MCP alternative: `project_docs_bootstrap_plan`.

2. Inspect the plan JSON: `signals.files`, `signals.languages`,
   `signals.test_commands`, `signals.profile`, each planned
   `files[].action`, and `files[].preserved` (true means the existing file
   will be kept as-is; only `AGENTS.md`'s managed marker block refreshes in
   place). Heed `warnings` such as
   `family_docs_preserved`, and `sync_available` — they reroute work to
   `project-docs-optimize` or `--sync` instead of forcing writes.

3. Write files and repo profile metadata when the plan is acceptable:

   ```bash
   issueops project bootstrap --repo . --json
   ```

   Add `--sync` only for an explicitly requested refresh of existing
   generated docs.

4. Run the agent enrichment pass using `PROMPT.md` in this skill directory.
   It reads repository evidence and updates `.issueops` docs through the
   MCP contract (`project_docs_route` → `project_docs_read` →
   `project_docs_revise` / `project_docs_append`) rather than blindly
   accepting static template text. It also evaluates the engineering
   standards catalog at `references/engineering-standards.md` in this skill
   directory (layered/hexagonal/onion/clean architecture, DDD, SOLID, OOP,
   clean code, error/exception handling, OpenAPI/Swagger, testing best
   practices, and adjacent topics) and maps every evidence-confirmed topic
   to its single owning document per the catalog's topic-to-doc map.
   The MCP workspace tools take an `authority_file` from
   `issueops mcp authorize --workspace-root "$PWD" --json`, issued with the same
   `ISSUEOPS_STATE_DIR` as the MCP service. When that path is unavailable,
   edit the docs directly as `PROMPT.md` allows and report that the MCP path
   was not used.

5. Verify:

   - `AGENTS.md` contains the behavioral top block and only the managed
     marker block was added or updated;
   - every expected root index, module starter, single-owner root, and
     `documentation/manifest.json` exists;
   - each family root links into its module directory and each module
     starter back-links its root;
   - the MCP maintenance path works: `project_docs_route`,
     `project_docs_read`, `project_docs_revise`, `project_docs_append`;
   - the project-docs-optimize checker reports zero violations on the fresh
     layout when its script is available:

     ```bash
     uv run --directory <issueops>/skills/project-docs-optimize \
       python -m scripts.check --root "$PWD" --mode check --json
     ```

   - the smallest relevant repo validation command still passes when docs
     make claims about workflow or commands.

## Generated Documents

`AGENTS.md` (behavioral block + managed marker block) and the
`.issueops` family: ARCHITECTURE, CAUTIONS, COMMIT_POLICY,
CONSTITUTION, CONVENTIONS, TECH_STACK, TESTING, OPEN_API_SPEC, ADR,
OPERATIONS, AGENT_WORKFLOW. In client repositories (detected `frontend`
or `desktop-client` project type) bootstrap also creates
`.issueops/DESIGN.md`, a design-system contract that points at a
curated root `DESIGN.md` when the repo already maintains one.

## Completion Evidence

Report: plan summary, files created vs preserved vs synced, enrichment
updates with their evidence, verification command results, and remaining
unknowns explicitly marked.

Files in this skill

  • PROMPT.md9.3 KB
  • SKILL.md6.8 KB
  • agents/openai.yaml358 B
  • references/engineering-standards.md13.9 KB

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.