Skill

dogfooding-issues

Processes open GitHub issues using self-development protocol, prioritizing research registry techniques with voting and testing. For dogfooding workflows, issue implementation, or research techniques.

Git

Github

Bash

git-workflow

npx claudepluginhub williamzujkowski/nexus-agents

Tool Access

This skill is limited to using the following tools:

ReadEditWriteBashGrepGlobTaskWebSearchWebFetch

Preview

<!-- CANONICAL SOURCES:

SKILL.md

Similar Skills

walk-the-issues

Grooms all open GitHub issues with subagents for clarity then loops implementation: branch, research, swarm code, test, commit, merge PRs until backlog cleared.

2 files

ux-toolkit

github-issue-workflow

174

Resolves GitHub issues via 8-phase workflow: fetch details, analyze requirements, implement solutions, verify correctness, code review, commit changes, create PRs. Activates on resolve, implement, fix requests or issue references.

4 files9 tools

developer-kit

team-collaboration-issue

682

Resolves GitHub issues through triage, root cause analysis, TDD, branch management, testing, and CI/CD-compliant pull requests. Ideal for team workflows on bugs and features.

1 file

rmyndharis-antigravity-skills

Stats

Stars10

Forks1

Last CommitMay 5, 2026

Actions

View Source View Plugin View on GitHub View README

Help us improve

Share bugs, ideas, or general feedback.

Dogfooding Issues Skill

Process open issues following the self-development protocol.

Full workflows:

Pre-Flight Checks

TZ='America/New_York' date '+%Y-%m-%d %H:%M:%S %Z'
git status --short
gh issue list --state open --limit 20

Issue Selection

Check Research Registry First

grep -A 5 "status: planned" docs/research/registry/techniques.yaml | \
  grep -B 5 "priority: P[12]"

Priority order:

P1 techniques with existing issues
P2 techniques with clear implementation path
Bug fixes blocking other work
Documentation gaps

Implementation

Follow CONTRIBUTION_GUIDE.md:

Research - Check registry, verify dependencies
Interface First - Define before implementation
TDD - Write failing test, implement, refactor
Registry Update - Change status: planned → status: implemented
Commit - Reference issue, update registry

Quality Checklist

See CODING_STANDARDS.md:

Tests pass, lint/types clean
Coverage ≥ 80%
Registry updated (if technique)
Issue referenced in commit

Anti-rationalization — Dogfooding

Excuse	Counter
"Pick the easy issue first"	Easy + un-implemented is suspicious — usually means the issue isn't actually easy or the spec is wrong. Pick by impact, not by ease.
"Skip the research check"	The research registry is where prior thought lives. Skipping it means re-deriving solutions and missing prior decisions.
"Test coverage 80% is fine"	Per CLAUDE.md, the gate is 89.66% / 93.26% (statement / function). Don't ship below the existing floor.
"Implementation is partial — flag mark as 'in progress'"	Per implement-feature: don't mark implemented if partial. Either ship the slice or split the issue.

Red flags

Issue closed with PR linked but feature behind a flag
Research registry entry not updated to status: implemented
Test count dropped after the implementation (lost coverage)
Issue spawned >5 follow-up issues (scope creep — should have split)

Verification checklist

Research registry checked first
Tests pass at the existing coverage gate (89.66% / 93.26%)
Feature complete (not partial, not flag-gated unless flag is the design)
Research registry status updated to implemented
Issue closed with summary linking the PR