Skip to content

Issue Tracking

GitHub Issues are Nuzo's execution tracker.

Docs and roadmap explain direction. Issues should describe work that can be assigned, reviewed, and closed.

Labels

Use three label groups together when possible:

  • area:* for ownership;
  • type:* for the kind of work;
  • priority:* for sequencing.

Current area labels:

  • area:core
  • area:cli
  • area:mcp
  • area:codex
  • area:claude-code
  • area:docs
  • area:release

Current type labels:

  • type:feature
  • type:bug
  • type:docs
  • type:architecture

Current priority labels:

  • priority:p0
  • priority:p1
  • priority:p2

Use status:blocked only when an issue has a concrete external blocker.

Milestones

Current milestones:

  • MVP
  • Host Plugins
  • Lifecycle Hooks
  • Public Release

Milestones should stay small enough to guide execution. Do not use them as permanent categories.

Issue Quality

Good issues include:

  • a clear goal;
  • relevant package or docs boundaries;
  • concrete tasks;
  • acceptance criteria;
  • links to docs or prior decisions when useful.

Avoid issues that only say "improve docs" or "make plugin better". Convert them into a specific deliverable before starting work.

Periodic Issue Hunting

Run issue hunting after meaningful changes and before larger next-step work.

Check:

  • docs and README for stale product claims;
  • roadmap status against implemented packages;
  • CLI, MCP, and plugin commands shown in docs against real commands;
  • generated artifacts or runtime memory accidentally staged;
  • GitHub Issues for duplicates or completed work that should be closed;
  • GitHub Pages and release workflow status when public docs change;
  • tests and validation gaps around recently changed behavior.

Small safe fixes can be handled immediately. Larger work should become a focused GitHub Issue with labels, milestone, concrete tasks, and acceptance criteria.

Security

Issues must not include:

  • runtime memory files;
  • real memory exports;
  • API tokens;
  • credentials;
  • cookies;
  • private keys;
  • private user data.

Use fake data for examples and reproduction steps.