agentleFS
AbilityaiAbilityai/cornelius/.claude/skills/scheduled-run/SKILL.md

scheduled-run

Skill109 starsChanged 39 days agoMITView on GitHub

Wrapper for scheduled playbooks - handles git sync before and after execution

  • Commits and pushes

The file

We split this skill into the rules it gives an agent

3 rules. About 0 are common, 3 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    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:

    All changes committed and pushed after running

    line 14
  2. Distinctive:

    If conflicts, abort and log error.

    line 39
  3. Distinctive:

    If no changes, skip to completion.

    line 67

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 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

Scheduled Run

Wrapper that runs any playbook with git sync. Use this for all scheduled executions to ensure:

  • Latest changes pulled before running
  • All changes committed and pushed after running

Purpose

Ensures scheduled playbooks on Trinity stay in sync with the GitHub repository.

State Dependencies

SourceLocationReadWriteDescription
Git repo.git/✓✓Pull before, commit+push after
Playbook outputvaries✓Whatever the playbook produces

Arguments

$ARGUMENTS = The playbook name to run (e.g., refresh-index, coherence-sweep)

Process

Step 1: Pull Latest Changes

git pull --rebase

If conflicts, abort and log error. Do NOT run the playbook on a dirty tree.

Step 2: Run the Playbook

Scheduled playbooks run at the core read-scope by default. BRAIN_READ_SCOPE is unset here, so every run_search.sh / run_connections.sh call inside the playbook fails closed to core (the fingerprint) - which is the intended scope for incubation-loop, refresh-index, coherence-sweep, gjopen-refresh, and ai-crystallize. Perception/cross-domain playbooks that legitimately need non-core content (e.g. domain-watch) set a wider BRAIN_READ_SCOPE=... inline on their own search commands; do not widen it here. (Setting export BRAIN_READ_SCOPE=core here would be a redundant fail-safe - the default is already core - and would not propagate across the separate Bash invocations a playbook makes anyway.)

Invoke the specified playbook using the Skill tool:

Skill: $ARGUMENTS

Wait for completion.

Step 3: Check for Changes

git status --porcelain

If no changes, skip to completion.

Step 4: Commit Changes

git add -A && git commit -m "$(cat <<'EOF'
Scheduled: $ARGUMENTS $(date '+%Y-%m-%d %H:%M')

Automated execution via /scheduled-run

Co-Authored-By: Claude <noreply@anthropic.com>
EOF
)"

Step 5: Push to Remote

git push

Error Handling

ErrorRecovery
Pull fails (conflicts)Abort, log error, do not run playbook
Playbook failsStill attempt commit/push of partial changes
Push failsLog error, changes remain local

Autonomy Guide

What should be scheduled (mechanical, no creative decisions):

ScheduleCommandCron (UTC)TimeoutPurpose
Incubation Loop/scheduled-run incubation-loop0 * * * *45mHourly thinking on active topics (self-seeds via starvation floor when empty) — the :00 anchor
Domain Watch/scheduled-run domain-watch30 */2 * * *30mEvery 2h at :30 — perception layer; auto-activates topics (incl. starvation floor)
GJ Open Refresh/scheduled-run gjopen-refresh30 1 * * *30m01:30 daily — external forecasting data [disabled]
Refresh Index/scheduled-run refresh-index30 3 * * *40m03:30 daily — FAISS + BDG bootstrap
Coherence Sweep/scheduled-run coherence-sweep30 5 * * 130mMon 05:30 — staleness/health briefing [disabled]
AI Crystallize/scheduled-run ai-crystallize30 7 * * 130mMon 07:30 — AI synthesis of converged topics → AI Crystallizations/ (ai-inferred)
Git Sync(raw git)45 */6 * * *10m:45 every 6h — redundant with scheduled-run git hygiene [disabled]

Spacing rule: incubation holds the hourly :00 slot; every other job runs at :30 of an odd hour (domain-watch only runs even hours), so no two jobs start in the same minute and their git commits never collide. Timeouts match each skill's work budget (the 15-min platform default is too short for FAISS rebuilds and multi-domain scans).

What should NOT be scheduled (requires human judgment):

SkillWhy manual
/auto-discoveryGenerates new connections - that's intellectual work
/deep-researchCreates new content from external sources
/detect-tensionsProductive contradictions need human interpretation
/compute-lifecyclePromotions to framework status need human decision
/integrate-recent-notesIntegration choices are creative decisions
/analyze-kbUseful on-demand, not worth weekly automation

The principle: Automate measurement and maintenance. Never automate creation or judgment into the endorsed knowledge base. The refinement (2026-06): autonomous synthesis is permitted when its output is segregated and tagged ai-inferred - the starvation floor self-seeds topics, and ai-crystallize writes AI conclusions to Brain/05-Meta/AI Crystallizations/ (never 02-Permanent/). The gate that stays human is the endorsement act: only the user promotes an ai-inferred conclusion into permanent, endorsed knowledge via /manage-thinking-topics crystallize.

The coherence sweep is the one weekly artifact that earns automation - it's a Monday briefing that says "here's what changed, here's what might be stale, here's what's evolving." You glance at it, act on what matters, ignore the rest.

Completion Checklist

  • Git pull completed successfully
  • Playbook executed
  • Changes detected and committed (if any)
  • Changes pushed to remote

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.