7
GitHub Stars
1
Bundled Files
2 months ago
Catalog Refreshed
4 months ago
First Indexed
Readme & install
Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.
Installation
Preview and clipboard use veilstrat where the catalogue uses aiagentskills.
npx veilstrat add skill lagz0ne/c3-skill --skill c3-onboard- SKILL.md13.6 KB
Overview
This skill bootstraps C3 (Context-Container-Component) architecture documentation from scratch using a staged Socratic discovery workflow. It runs a strict 3-stage process (Inventory → Details → Finalize) to create a complete .c3 directory with ADR, context, container, component, and ref files. It is designed for projects that have never had C3 docs and will not modify existing .c3 assets unless you explicitly re-onboard.
How this skill works
The skill first inventories the codebase and creates ADR-000 to capture context, containers, components, abstract constraints, and candidate refs. After a discovery gate, it generates container README files, one component file per component, and curated ref files following templates. Finally it runs integrity checks and routes to c3-audit to validate the new docs before marking ADR-000 implemented. All user questions use AskUserQuestion unless denied, in which case Assumption Mode applies and decisions are marked.
When to use it
- You want to adopt C3 for a project that has no .c3/ directory yet (fresh project).
- You need to scaffold architecture docs: create .c3/, containers, components, refs, and ADR-000.
- You are starting C3 onboarding for the first time or bootstrapping a new codebase.
- You want a guided, gated workflow to discover and document context, boundaries, and responsibilities.
- You need a single-file-per-component and per-container-directory C3 layout created from templates.
Best practices
- Do not run on repos where .c3/ already exists; use c3-audit, c3-change, or c3-query instead.
- Always answer AskUserQuestion when prompted; if denied, Assumption Mode is enabled and assumptions are recorded.
- Follow the non-negotiable file rules: one file per component, one directory per container, refs in refs/.
- Complete Inventory fully before creating container/component files to avoid partial docs detection.
- Extract cross-cutting patterns into refs when content would remain stable across technology swaps.
Example use cases
- Initialize C3 for a newly created monorepo with web and api folders to produce .c3/ and ADR-000.
- Scaffold architecture documentation for a legacy codebase that has never been documented with C3.
- Onboard a project by discovering containers and components, then create container READMEs and component files.
- Generate a mermaid overview diagram, abstract constraints table, and a clear list of code references per component.
FAQ
Assumption Mode is set for the session: no further questions are asked, assumptions are documented in ADR-000 (marked [ASSUMED]), and low-impact choices proceed automatically.
Can this skill update existing .c3 docs?
No. This skill is for first-time onboarding only. For updates or changes to existing docs use c3-change or c3-audit.