asdecided/core/v0-17-x-adoption

Claude Code and agent skill by asdecided/core

Trust tier 3
  • community
Source check pending

Checking that the SKILL.md file is still in the repository.

Install

npx vskill@latest install asdecided/core/v0-17-x-adoption

SKILL.md

v0.17.0 — Single-Document Import Skill

Status

Achieved

Context

A common adoption blocker is the on-ramp: new users have existing decisions in Confluence, Notion, or loose Markdown and no guided way to get one into RAC format. The rac-ingest skill converts and can mint identity across a whole directory, but it is broad and unguided. This milestone adds a focused, human-in-the-loop sibling for the most common first step — turning one existing document into one valid artifact — opening the v0.17.x adoption series (the import on-ramp, per-type standards, and CI distribution).

Outcomes

  • An agent in a host project has a bundled procedure to reshape exactly one

existing document into one valid RAC artifact, with an explicit human-review step before any file is written and rac validate as the deterministic close.

  • The skill adds no AI to RAC core: it proposes, the human ratifies, and the

deterministic CLI checks. It reuses the released command surface only.

  • The bundled skill set grows to four without weakening the install or

byte-equality contracts.

Initiatives

  • Bundle rac-import alongside the existing skills, packaged identically

(resource under rac.skills, byte-identical dogfood copy under .claude/skills/, per-skill sync test). Its procedure: take one document (paste or path); convert non-Markdown via rac ingest; choose the type and read its real shape with rac schema <type>; draft only from source content, flagging required sections the source does not cover; present the draft and require the user to confirm the type, title, and any relationships before writing; scaffold with rac new (which mints the id); then close on rac validate, offering fixes until it exits 0.

  • Single-document by scope: the skill refuses a directory, a wiki, or several

documents and points to rac-ingest for broader conversion. Relationships are suggestions the source itself names, each confirmed with rac resolve / rac find — never inferred by scanning the repository.

  • Describe released CLI behaviour only: rac ingest, rac schema, rac inspect,

rac new, rac validate, rac resolve, rac find. Normative-language conventions (BCP-14) are authoring guidance, not a check the skill claims, because rac validate does not enforce them yet.

  • Document the on-ramp in the README ("Importing an existing decision") and

refresh the skill list/install goldens for the new entry.

Success Measures

  • rac skill install rac-import writes .claude/skills/rac-import/SKILL.md;

rac skill list enumerates it; re-running install exits 1 and writes nothing.

  • The packaged skill is byte-identical to its dogfood copy (enforced by test).
  • Following the skill, three realistic messy source documents (a decision, a

requirement, a design) each produce an artifact that passes rac validate, faithful to the source with no invented facts.

  • pytest, rac validate rac/, and rac relationships rac/ --validate pass.

Assumptions

  • Project-level skills (.claude/skills/) remain the documented discovery path.
  • The host project has the rac CLI installed; the skill does not manage install.

Risks

  • Overlap with rac-ingest could confuse users. Mitigated: the descriptions and

the skill body state the single-document, human-reviewed scope and cross-refer rac-ingest for broader conversion.

  • Skill prose drifts from CLI behaviour as commands evolve; the sync tests catch

drift between copies but not against the CLI itself.

Related Decisions

  • adr-003
  • adr-007
  • adr-021

Related Requirements

  • rac-growth-agent-skill
  • rac-growth-adoption

Related Roadmaps

  • v0.10.5-review-and-ingest-skills
  • v0.10.5-bundled-agent-skill

Security

  • Security scanPassed, 95 of 100Pattern checks on the skill's files.
  • Model reviewPassed

Quality evaluation

  • Task evaluationNot runNo evaluation yet. Run one in Skill Studio.

Source

Popularity

7-day trend score
5

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
vskill VERIFIED badge
[![vskill VERIFIED](https://verified-skill.com/api/v1/skills/asdecided/core/v0-17-x-adoption/badge)](https://verified-skill.com/skills/asdecided/core/v0-17-x-adoption)

Something wrong with this skill or its page? Tell us.