agentleFS
zauberzeugzauberzeug/game-of-cards/.claude/skills/tune-cadence/SKILL.md

tune-cadence

Skill7 starsChanged 6 months agoMITView on GitHub

Query or change how often this repo's autonomous GitHub Actions run (pull-card, audit-deck, refine-deck). AUTO-INVOKE when the user asks what the current cadence is, how often pull-card/audit/refine run, to speed up or slow down the deck, to run pull-card every N hours, to change the autonomous cadence, or to retune the workflows. Repo-local dev skill that wraps scripts/set_cadence.py; not a packaged goc skill.

  • Commits and pushes

The file

We split this skill into the rules it gives an agent

1 rule. About 0 are common, 1 is distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Follow AGENTS.md "Parallel-Agent Commit Safety" — stage explicit paths, verify git diff --cached, commit with a pathspec.

    line 63

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Tune the autonomous workflow cadence

This repo drives three autonomous agents from GitHub Actions:

  • pull-card (.github/workflows/pull-card.yml) — drains the queue.
  • audit-deck (.github/workflows/audit-deck.yml) — files one new card per run.
  • refine-deck (.github/workflows/refine-deck.yml) — one deck-hygiene pass per run.

Their schedules are GitHub Actions cron: literals, which can't read a config file. scripts/set_cadence.py rewrites those literals (and the managed # cadence: comment beside each) from a small interval spec, so this skill is just a thin wrapper around it.

Query the current cadence

python3 scripts/set_cadence.py --show

Prints each workflow's cron plus its human-readable interval.

Change the cadence

python3 scripts/set_cadence.py --pull 1h --audit 3h --refine 3h
  • Pass only the workflows you want to change; omit the rest to leave them.
  • Interval specs: <N>h where N divides 24 (1, 2, 3, 4, 6, 8, 12) or 24h; <N>d for every-N-days; and 1w for exact weekly. Note Nd (N≥2) becomes a day-of-month */N cron step that realigns each month, so it's "roughly every N days" — the gap across a month boundary is shorter. 1w is exact (day-of-week, every Monday); there is no clean cron for every-N-days or every-N-weeks.
  • An optional +P hour-phase suffix shifts a schedule by whole hours — use it to keep two workflows on the same interval a full hour apart (the minute offsets below only separate them within an hour slot). 3h+1 fires at 01:xx, 04:xx, … instead of 00:xx, 03:xx, …. For Nh (N≥2) the phase must be < N; for 24h/Nd/1w it sets the hour-of-day (0–23); 1h takes no phase.
  • Minute offsets are fixed (pull :13, audit :15, refine :45) so the three deck-mutating agents never launch on the same minute and race on main.

Make it take effect

Scheduled workflows run from the default branch, so a cadence change does nothing until it is committed and pushed to main:

git add .github/workflows/pull-card.yml \
        .github/workflows/audit-deck.yml \
        .github/workflows/refine-deck.yml
git commit -m "ci: retune autonomous cadence"
git push

Follow AGENTS.md "Parallel-Agent Commit Safety" — stage explicit paths, verify git diff --cached, commit with a pathspec.

Cost note

Each run of any of these workflows is a full Opus agent under bypassPermissions. Faster cadence means more token spend; dial it back toward daily when spare-token headroom is tight. The cap on pull-card's per-tick self-trigger chain is separate — it lives in MAX_ITERATIONS inside pull-card.yml, not here.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.