菜单
Persistent notes a scheduled/recurring agent reads as its FIRST action and writes as its LAST action, so a weekly/nightly autonomous run "knows where it left off." Use inside /loop or /schedule tasks that span multiple wake-ups. Pairs with hypothesis calibration and the experiment queue.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Pre-commit anti-slop audit — checks staged/dirty files for code slop (debug artifacts, single-use helpers, dead imports) AND documentation slop (stale claims, broken internal links, new .md files, redundant sections). Writes a sentinel so the pre-commit-gate hook lets the commit through. MUST be invoked before `git commit` in this repo. Triggers when the agent is about to commit, when user says "commit", "clean", "ready to commit", or when the pre-commit-gate hook blocks.
Pull expert skills into this harness from the Rockie platform catalog (~300 skills across ML training/inference, biology, chemistry, physics, databases, coding). Use BEFORE writing domain guidance from scratch or fumbling an unfamiliar framework — if the task names a library (vLLM, GRPO, TRL, AlphaFold, DuckDB, LAMMPS), check the catalog first. Browse `rockie skill catalog --search X --json`, pull into `.claude/skills/`, invoke immediately. Silently no-ops when the Rockie CLI is absent or logged out.
Canonical around-the-clock research loop. Defines the agent's outer loop — read taste corpus + queue, pick the next experiment, mutate the explicitly-declared mutation surface, run the experiment under a hard time budget against a frozen metric, score, codify, repeat. Augmented with Karpathy's sharp primitives (frozen metric, time cap, explicit mutation surface) and a sustained-campaign layer (concurrent run/plan/write-up pipeline, verdict protocol, novelty re-verification gate) for multi-day operations.
Build any goal-declared agent end to end. Turns one declared goal into a portable, self-hardening, publishable agent repo on the Claude-Code harness. Declare goal, pick runtime/tools and write policies/hooks/safeguards, scaffold a portable repo (config-over-code split), build, run a fresh no-memory adversarial gauntlet, iterate with the clean-pass counter resetting on any failed round, publish ONLY after passing twice with zero CRITICALs, then export a .af agent-file. Biased to emit agents that contain their own internal adversarial critic loop. Triggers on "build an agent", "make me an agent that", "agent builder", "scaffold an agent", "publish this agent", "export to .af", "/agent-builder".
Lab composer command scaffold for `/build-agent <name>`. In Agent Builder A1 this only creates a lab-scoped draft agent record and opens that lab's agent detail route for editing. It does not build, run, evaluate, observe, or deploy agents yet.
Narrow quickstart wrapper for pre-baked Deploy-card requests. Triggers on server-validated structured quickstart deploy prompts such as "Quickstart deploy request: track: deploy model: <id> source: quickstart-picker quickstart_intent_id: <id>" and explicit "deploy this featured model" requests from Rockie's quickstart picker when validation metadata is present.
| name | scheduled-notes |
| description | Persistent notes a scheduled/recurring agent reads as its FIRST action and writes as its LAST action, so a weekly/nightly autonomous run "knows where it left off." Use inside /loop or /schedule tasks that span multiple wake-ups. Pairs with hypothesis calibration and the experiment queue. |
Ported pattern: Cognition's "Devin can now schedule Devins" — the weekly task hands itself a note. Without this, a scheduled run starts cold every time and can't build on its own prior conclusions.
At the project root, maintain SCHEDULED_NOTES.md. It's a single
bounded-length markdown file that the scheduled agent edits in place.
Suggested structure:
# SCHEDULED_NOTES.md
_Last updated: <ISO timestamp>_
_Current stage: creative_
## Open threads (pick these up next)
- Line 1: one-sentence thread + pointer to relevant journal node
- Line 2: ...
## Recent conclusions (compact)
- YYYY-MM-DD: outer-product init reduces warmup loss 0.02 (closed #42)
- ...
## Blockers / decisions needed
- (empty or a short note if human input would help)
Keep the whole file under ~2 KB. When it grows, compact into a line of "conclusions" and drop the oldest threads.
First action of every scheduled run:
cat SCHEDULED_NOTES.md
Last action before exit:
# Update timestamp + current stage + open threads + recent conclusions
# Use apply_patch.py or Edit — either is fine.
The natural composition is:
SCHEDULED_NOTES.md → know where you left off.python3 .claude/scripts/queue.py next --json → pop next queue item.SCHEDULED_NOTES.md (what got done, what's next).If the agent stops writing the notes, you lose continuity. The
check-scheduled-notes helper (below) can flag when the file's
Last updated: timestamp drifts from reality.
# Optional: add to an hourly /loop. Portable mtime in seconds-since-epoch
# (works on GNU coreutils and BSD/macOS).
mtime() {
if stat -c %Y "$1" 2>/dev/null; then :;
else stat -f %m "$1" 2>/dev/null; fi
}
if [ "$(date +%s)" -gt $(( $(mtime SCHEDULED_NOTES.md) + 86400 )) ]; then
.claude/scripts/notify.sh 2 "scheduled-notes stale" \
"SCHEDULED_NOTES.md not updated in >24h; something is wrong"
fi
STATE.md is Claude's always-on context (it's referenced by CLAUDE.md).
SCHEDULED_NOTES.md is specifically for the agent-to-its-future-self
pattern. Keeping them separate keeps STATE.md clean for human reading
and lets scheduled runs churn through notes aggressively.