Quickstart - Stet

This first session verifies setup and creates an onboarding receipt. It stops before a model smoke, probe, or rules evaluation.

1. Install the CLI

macOS and Linux:

curl -fsSL https://raw.githubusercontent.com/Stet-AI/stet-cli/main/install.sh | sh
stet --version

Windows PowerShell:

$installer = Join-Path $env:TEMP "stet-install.ps1"
Invoke-WebRequest -Uri https://raw.githubusercontent.com/Stet-AI/stet-cli/main/install.ps1 -OutFile $installer
try { & $installer } finally { Remove-Item -Force $installer }
stet --version

The installer scripts and agent skill come from the current public Stet-AI/stet-cli/main source. The installers select the latest stable CLI by default; verify the installed binary with stet --version. Commercial Stet workflows require local authentication:

stet auth login
stet auth status

Install the agent skill separately:

npx skills add https://github.com/Stet-AI/stet-cli.git --skill stet --all
npx skills list

The CLI runs evaluations and manages artifacts. The skill tells your coding agent how to choose a workflow and read the canonical result.

3. Verify prerequisites

Run the checks that match the backend you plan to use:

gh auth status
stet --version
stet auth status
npx skills list
docker info
command -v uv

The default isolated path needs Docker, Python 3.12 or newer, uv, and authentication for the model provider you eventually plan to evaluate. GitHub CLI access is needed for private release overrides or PR-backed discovery.

4. Ask your agent to onboard one repository

From the repository you want to measure, ask:

Use the Stet skill. Onboard this repo for Stet evals. Ask what work I want
Stet to track, read CI first, and choose the real test command. Inspect merged
PRs/commits for representative work (use read-only subagents for this
investigation when available), create the Harbor setup, and build a
representative starter dataset. Choose and pass the narrowest credible bounded
verifier to `stet suite build` with `--test` (alongside the manifest and a
bounded `--workers` value). Treat `bazel test //...`, unfiltered `pytest`,
`go test ./...`, and workspace-wide scripts as broad; do not launch one without
my explicit approval. If no credible bounded verifier exists, stop before
building and propose alternatives. Do not broaden merely to increase yield.
Report scope, skipped tasks and reasons, then stop before any model eval. Treat
Docker or replay checks as setup validation, not a model eval.
Wait for `build-summary.json` before treating the build as complete, then return
the onboarding receipt with the selected slice, scope, skips, and confidence.

Your agent should inspect CI and build files, ask about the product areas and work mix you care about, select real merged work, and create the repository’s Stet harness files. A starter slice should cover the work you intend to track; it is not automatically representative just because it was recent. stet suite build may use model-assisted install_config synthesis by default to improve verifier fidelity. If this onboarding must spend no model calls, have the agent pass --llm-install-config=false to the suite build and record that choice in the receipt; the no-model path can produce lower-fidelity broad verifiers.

5. Inspect the onboarding receipt

Ask your agent to summarize the receipt before doing more work:

Use the Stet skill. Read the Stet onboarding receipt. Summarize the
candidate-task funnel, selected starter slice, representativeness rationale,
subsystem/path coverage, difficulty mix, skipped-task reasons, Docker/test
setup validation, confidence, and recommended next step. Do not launch more
work yet.

If setup or task selection is blocked, use Troubleshooting. When the receipt is credible, choose a bounded workflow from Choose a workflow. Do not treat this setup check as a model result. To expand a credible starter slice, ask for a bounded follow-up:

Use the Stet skill. Expand the existing onboarding slice by one bounded
subsystem or task family. Read the current receipt first, preserve its test
command with an explicit `--test`, and explain the expected coverage gain.
Use read-only investigation where useful; do not broaden just to increase
yield. Stop before building if no narrow verifier is credible, and propose
alternatives. Report added scope and skips, then stop before any model eval.

​1. Install the CLI

​2. Sign in and install the skill

​3. Verify prerequisites

​4. Ask your agent to onboard one repository

​5. Inspect the onboarding receipt

1. Install the CLI

2. Sign in and install the skill

3. Verify prerequisites

4. Ask your agent to onboard one repository

5. Inspect the onboarding receipt