kdeldycke/repomatic-topics

Context

!grep -A1 '^\[project\]' pyproject.toml 2>/dev/null | tail -1 !gh api repos/{owner}/{repo} --jq '.topics | join(", ")' 2>/dev/null || echo "NO_GH_ACCESS"

Instructions

You help users optimize their GitHub repository topics (tags) for maximum discoverability, balancing accuracy with competitive positioning on GitHub topic pages.

Determine repository

• Read pyproject.toml for the project name, description, and keywords list.
• Use gh api to fetch the current GitHub topics from the repo.

Argument handling

• audit (default when $ARGUMENTS is empty): Analyze current topics and recommend changes. Do not apply.
• apply: Analyze, confirm with the user, then apply via gh api --method PUT repos/{owner}/{repo}/topics.

Analysis process

1. Inventory features. Read the codebase (CLI commands, workflows, config options) to understand what the project actually does.

2. Assess current topics. For each existing topic, fetch https://github.com/topics/{topic} and evaluate:

   • How many repos use the topic.
   • What star count is needed to appear on page 1 (above the fold).
   • Whether this repo can realistically rank there.

3. Classify competitiveness. For each topic:

   • High competition (page 1 requires 10k+ stars): effectively invisible. Drop unless the topic is core identity.
   • Medium (page 1 requires 1k-10k stars): keep only if the repo is borderline viable.
   • Low/niche (page 1 requires <1k stars or few total repos): high ROI — keep or add.

4. Propose candidate topics. Search for niche topics where the repo could rank in the top 3. Look for:

   • Feature-specific topics (e.g., github-labels, reusable-workflows) over generic ones (e.g., labels, ci-cd).
   • Compound topics that narrow the audience (e.g., python-automation vs automation).
   • Topics with few repos where claiming them gives instant #1 ranking.

5. Present recommendations. Show a table with:

   • Topics to keep (with rank assessment).
   • Topics to drop (with reason: too competitive, inaccurate, etc.).
   • Topics to add (with repo count and expected rank).

GitHub allows up to 20 topics. Fill all 20 slots — unused slots are wasted discoverability.

Required topics for awesome lists

If the repository name starts with awesome-, it must have both the awesome and awesome-list topics. This is enforced by awesome-lint. Flag any missing required topic during audit, and add them during apply.

After analysis

• Confirm changes with the user before applying.
• When applying, update both GitHub topics (gh api --method PUT repos/{owner}/{repo}/topics) and pyproject.toml keywords to stay in sync.
• The pyproject.toml keywords serve PyPI search — keep them aligned with GitHub topics.

Next steps

Suggest the user run:

• /repomatic-lint to check for any issues introduced.