Skip to content
Get packs

GitHub Cache Hygiene

Check GitHub issues and PRs cheaply via local archive first, then live gh reads with rate-limit care

Source Download skill
Repository
steipete/agent-scripts
Stars
7,433
Licence
MIT
Files
2 files

What it does

When installed, the agent searches a local Gitcrawl archive of a repository's issues and pull requests first, which is cheap and avoids GitHub API calls. It only uses the GitHub CLI (gh) for live data when current state is needed, such as before a write, merge, or CI check, and it spaces out CI polling and reuses earlier output to avoid hitting rate limits. It is written for a specific setup where gh is routed through a caching shim called Octopool.

Great for

  • Maintainers or developers who triage many issues and PRs and are hitting GitHub API rate limits
  • Teams running the Gitcrawl archive and Octopool shim who want consistent gh usage rules for agents
  • Anyone who wants an agent to check CI status after a push without polling excessively

Skip it if

  • Users without Gitcrawl or Octopool installed: much of the guidance depends on those tools and will not apply, and the skill does not explain how to install them
  • Situations needing guaranteed up-to-the-second accuracy from cached data: the skill itself warns archive state can lag GitHub, so live calls are still required for decisions
  • Anyone wanting it to bypass or change authentication: the skill forbids running octopool login or altering tokens, auth, PATH or config, so it is not a tool for changing credentials

Try asking

  • “Search this repo's open PRs about the login bug in the local archive, then confirm the current state with gh before I merge anything”
  • “Check whether an issue about dependency upgrades already exists in owner/repo before I open a new one”
  • “Watch the CI run for my latest push at measured intervals and only fetch logs if it fails”
  • “Show my Octopool cache hit rate for the last 24 hours to see whether rate limits are caused by my agent”
Step-by-step guide: how to use GitHub Cache HygieneInstall, first run, examples, tips and FAQs

Install in 3 steps

  1. 1

    Download the skill

    The ZIP contains the 2 original files, the MIT licence and an attribution note.

    github-cache-hygiene.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

    mkdir -p ~/.claude/skills && unzip github-cache-hygiene.zip -d ~/.claude/skills/

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip github-cache-hygiene.zip -d .claude/skills/

    Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.

  3. 3

    Ask for the task

    No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.

Get better results

  • Ask for a specific repo and search terms so the first archive read stays narrow and the agent does not loop over every result
  • Request live state explicitly (for example 'latest state') when a decision depends on it, since archive results can be stale
  • Check that the shim is actually in use on your machine before relying on the cache numbers, because the skill's guidance assumes Peter's Octopool setup

What's inside

  • agents/openai.yaml
  • SKILL.md

Redistributed unmodified from steipete/agent-scripts under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.