Agtmls New Skill — Development skill for Claude Code
Scaffold a new AgtMLS skill with its metadata and both eval cases.
How to install Agtmls New Skill
Installs to ~/.claude/skills/sebastienrousseau-agtmls-agtmls-new-skill/SKILL.md
mkdir -p ~/.claude/skills/sebastienrousseau-agtmls-agtmls-new-skill && curl -fsSL https://raw.githubusercontent.com/sebastienrousseau/agtmls/HEAD/commands/agtmls-new-skill.md -o ~/.claude/skills/sebastienrousseau-agtmls-agtmls-new-skill/SKILL.md Restart Claude Code, or start a new session, for it to be picked up.
What Agtmls New Skill does
description: Scaffold a new AgtMLS skill with its metadata and both eval cases. license: MIT
Scaffold a skill, then author it against the contract.
python3 scripts/agtmls.py scaffold-skill
python3 scripts/agtmls.py scaffold-skill --bundle
That writes `skills//` with `SKILL.md`, `reference.md`, `metadata.json`, and both eval cases.
Write the description first — it is the entire routing surface, and the tokenizer does not stem, so use the exact wo
Alternatives in Development
- HERO Anti OverDefense — HERO = Hashing · Edge cases · Rubrics · Overbuild — the four shapes coding agents over-defend in 333 ★
- Eval Author — Draft a new eval case (eval/cases/ .json) + scaffold its golden folder, golden_pending until captured on the V 136 ★
- Generate Ddd — Generate DDD components 63 ★
Full documentation available on GitHub
View Source RepositoryRelated Skills
New Client
Onboard a client — scaffold a profile in the working project and seed its eval cases.
Architect Prove
Run the prove stage. Writes at least ten eval cases including a refusal and an injection case, runs the harnes
Eval Cases
Grow an eval case set from the placeholder rows to a real 15-25 - harvest genuine cases out of the repository
Eval Init
Scaffold an eval from the bundled eval template into the repo's evals/ project (shared _env.py and cache once
Agtmls Release
Walk the guarded AgtMLS release flow from version bump to published artifacts.
Agtmls
Run the local AgtMLS registry health check.
Related Agents
Scaffold Inspect
Sets up inspect-ai evaluation configurations for all runs in a designed experiment. Reads experiment_summary.y
Metadata
Registers or updates all metadata in the Data Basis backend — dataset, tables, columns, observation levels, co
Docs Link Checker
Validates both external and internal links in documentation files to ensure they are not broken. Maintains a c