Quickstart

Skillet uses the coding-agent CLI you already run. The Skillet CLI manages files, validates contracts, and runs evals; your agent writes the specification, instructions, and cases.

• Node.js 20.11 or newer
• Codex, Claude Code, or another agent CLI for eval runs

npx -y @sentry/skillet@latest init

With pnpm, use pnpx @sentry/skillet@latest init. The authoring skill uses the same explicit @latest package command so its CLI-provided formats and guidance do not drift behind a global installation.

init asks before installing the skillet-authoring skill through dotagents in user scope under ~/.agents.

If you prefer a global binary, install it explicitly:

npm install -g @sentry/skillet
skillet init

Installed binaries check npm at most once per hour and recommend the current package command when an update is available.

Install the authoring skill in user scope directly:

npx -y @sentry/dotagents@latest --user add getsentry/skillet skillet-authoring

The add command records the dependency in ~/.agents/agents.toml and installs it immediately for supported agents. No separate install command is required.

Run install later to refresh declared skills:

npx -y @sentry/dotagents@latest --user install

Or tell your agent:

Install the skillet-authoring skill for me.

In your coding agent, ask:

Create a skill that enforces our commit conventions.

The authoring skill will:

1. Create the skill directory.
2. Clarify ambiguous behavior.
3. Write and validate spec.md.
4. Render SKILL.md.
5. Add eval cases.
6. Run validation and evals.

Full evals and baselines start authenticated agent CLI sessions. They can take time and consume model usage. Built-in harnesses use low effort and run up to four trials or skill/baseline variants within one case in parallel. Baseline remains opt-in.

From the new skill directory:

npx -y @sentry/skillet@latest status
npx -y @sentry/skillet@latest validate
npx -y @sentry/skillet@latest eval --dry
npx -y @sentry/skillet@latest eval --baseline

status reports the next step, validate checks the complete contract, eval --dry finds checks that pass before the agent runs, and --baseline compares pass rates with and without the skill.

1. Follow Create Your First Skill for the complete artifact flow.
2. Read Specifications before changing skill behavior.
3. Read Write Agent Instructions and Write Honest Evals before editing derived files.
4. Read Understand Eval Results before interpreting trials, baselines, or lift.