Init spec
Personal agent skills for Claude Code, Cursor, and other AI coding agents.
npx -y skills add guillevc/skills --skill init-specAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
One thing to look at
- 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
What its author says it does
Copied from the file, not written here
Bootstrap a living spec on an existing codebase by extracting the terms and decisions already baked into the code. Run once when adopting living-spec. User-invoked.
SKILL.md
1.4 KB, as published. Nobody here has run it
Init-spec
Cold-start the living spec on a repo that has none: reverse-engineer the vocabulary and decisions
already implicit in the code. You propose; record writes; the human ratifies. Runs once, at
adoption.
Steps
- Check it's needed. If a spec exists, stop and point the human at
recordandaudit. Check every homerecordwrites to:docs/GLOSSARY.md,docs/decisions/, and the agent docCLAUDE.md/AGENTS.mdfor an existing glossary or ADR section, its fallback home when nodocs/structure exists. - Survey. Read the code, README, and agent doc. Pull the ubiquitous language (recurring domain nouns) and the decisions baked in (datastore, protocols, hard limits, deliberate deviations).
- Filter hard. Keep only ADR-worthy decisions: hard to reverse, surprising, a real trade-off. Most code choices fail; drop them. Keep terms that are load-bearing vocabulary, not incidental.
- Propose. Present the draft: terms (one-line meaning) and decisions (1-3 sentence rationale). The human edits, cuts, approves.
- Write the approved set via
record.