菜单
Audits cron-scheduled skills for session mode, token waste, and cost efficiency — and enforces concise-reply constraints on cron contexts
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
| name | cron-hygiene |
| version | 1.0 |
| category | openclaw-native |
| description | Audits cron-scheduled skills for session mode, token waste, and cost efficiency — and enforces concise-reply constraints on cron contexts |
| stateful | true |
| cron | 0 9 * * 1 |
Cron jobs running in main session mode inherit the full conversation history — re-sending thousands of tokens of context on every invocation. A cron running every 5 minutes in main session mode can turn a $10/month setup into $80+ (issue #20092). This skill audits all configured cron skills weekly and enforces hygiene standards.
context-budget-guard monitors live session token usage in real-time. cron-hygiene is about the structural problem: cron jobs that are architecturally wasteful because they're configured in the wrong session mode.
Runs every Monday at 09:00 (cron: "0 9 * * 1"). On each wakeup:
main vs isolated)main mode × frequency × estimated context size| Rule | Check | Severity |
|---|---|---|
| Session isolation | Cron skills should use isolated session mode | High |
| Reply conciseness | Cron output should be < 500 tokens | Medium |
| Frequency sanity | Crons running < 10min apart need justification | Medium |
| Dead crons | Cron skills with no state update in 7+ days | Low |
| Overlapping crons | Two skills scheduled at the same time | Low |
Cron Hygiene Report — 2026-03-16
────────────────────────────────
5 crons audited | 2 violations | Est. monthly waste: $12.40
HIGH: persistent-memory-hygiene — runs in main session mode
Fix: add sessionMode: isolated to skill config
Est. savings: ~$8.20/month
MEDIUM: long-running-task-management — fires every 15min
Output not constrained (last run: 1,240 tokens)
Recommendation: add reply_token_limit: 200 to skill config
To fix a main-session cron manually:
python3 audit.py --fix-session <skill-name>
This writes sessionMode: isolated to the skill's OpenClaw config entry.
Handles PR review feedback by fetching comments, grouping issues, fixing one group at a time, and verifying before replies.
Reviews whether a skill will trigger reliably, guide useful behavior, avoid overlap, and produce testable outcomes.
Diagnoses OpenClaw provider, fallback, channel, MCP, and gateway config issues with read-only scans and stateful summaries.
Tracks required validation gates, records pass/fail/waived results, and reports readiness before task completion.
YAML-based delegation grant ledger — issues, validates, and tracks scoped permission grants for sub-agent expansions with token budgets and auto-expiry.
Walks the memory DAG to recall detailed context on demand — query, expand, and assemble cited answers from hierarchical summaries without re-reading raw transcripts.