releases of critical packages to allow community security review. Always review `pnpm-lock.yaml` changes in PRs.
## Coding Conventions
**General**:
- **Prefer Classes for Stateful Modules**: When encapsulating functionality that involves state (cached
VS Code extension that analyzes local AI session logs and surfaces insights in a webview dashboard. Read-only, zero telemetry, all analysis runs on the user's machine.
reload. The app is a long-lived menu-bar process, so **every code change requires a full rebuild and restart of the running app** to take effect — kill the running
pushes made while a PR is closed fire no events).
- Review-bot findings (pullfrog/Codex/Greptile) get verified against the code first, then either fixed with a reply + thread resolution, or refuted
within-allow, so do not create a
broad grant that defeats a required denial.
- Review Linux and macOS separately when a change affects enforcement,
networking, paths, processes, sockets, credentials
development server listens on port 8081 and proxies the backend on
port 8080.
```
## Generated code
Do not hand-edit generated files.
- Edit `api/v1/api.yaml`, then run `make api`.
- Edit `proto/**/*.proto
authoritative.** The frontend
> review rules it declares are read by the `claude-review`, `codex-review`,
> `code-review`, `fork-gpt-review`, and `fork-opus-review` workflows, and a
> `blocking: true` rule
browser/client code limited to public or anon keys. Server-only secrets
must stay server-side.
- Prefer project-specific `.env` files or a secrets manager for credentials.
## FINAL SELF-REVIEW
- Check
issues
- Close and update issues
- Search across issues
- **Pull Request Workflow**:
- Create pull requests
- Reviewcode changes
- Add comments to PRs
This integration enables Claude to participate as a full
which the markdown pipeline leaves
alone), every code fence gets a language, and no code block directly under an `##` heading -- say
what it is first.
## Adding things
| Goal | Action
cluster scenarios. Use existing synchronization primitives in the codebase rather than introducing new ones.
# CodeReviews
See [.github/code-review.md](.github/code-review.md) for codereview guidelines
Agents
**This file is the universal entry point.** Any AI coding agent (Claude Code, Codex, Hermes, Cursor, etc.) reads this to know which NotFair skill to run for which user