sluglist

Docs

Documentation

Everything is optional beyond one line of config. Start with the quick start; add pieces when you need them. Not sure which pieces? pick the scenario that matches you.

Quick start

Install the sluglist feedback widget and mount it with one line of config: a connector, and nothing else. ESM, CJS or a script tag from a CDN.

Capture modes, annotation & attachments

Full page, area, element and record-steps capture; annotation with arrow, box and text; error capture, the action trail, reporter form fields and attachments.

Connectors: deliver feedback anywhere

A connector is the only place that knows about storage. HttpConnector, the sluglist/contract validators for your endpoint, and recipes for Vercel Blob, S3/R2 and Supabase.

Checklist mode: structured acceptance

Pre-seed an acceptance checklist the client walks and checks off. Verdicts land in session.yaml as a coverage map — and an agent can generate the list from a branch diff.

Production & beta: presets, privacy, PII scrubbing

The beta and production presets: input masking, screenshot consent, PII text scrubbing, dismiss, localization, and the checklist to walk before real users see the widget.

Agents & CLI: the local feedback loop

Run npx sluglist dev, click feedback on your own app, and let an agent fix it — plus the four bundled skills, sluglist status, and the loop that runs until green.

Project conventions: .sluglist/PROJECT.md

One committed file holds your base branch, how to run and sign in, the hard limits QA must never complete, and evidence defaults — so the bundled skills stay upgradeable.

Artifact format (contract)

Every session is a folder: session.yaml plus one markdown file per issue with YAML frontmatter. A stable, additively-versioned contract for downstream parsers and agents.