run-frequency-guard

Set daily run caps by defining daily_max limits in memory/skill-caps.json or under skill_caps: in aeon.yml. The guard reads these configurations to enforce per-skill execution frequency.

Run frequency guardrail alerting detects overuse by counting today's executions logged in memory/token-usage.csv and comparing them against configured daily_max limits, sending a breach-only alert if caps are exceeded.

Yes, you can limit scheduled workflows by defining per-skill daily_max limits. The guard evaluates daily run activity and alerts you immediately when a capped skill exceeds its threshold.

Daily run limit breaches for autonomous operations are logged to memory/logs/${today}.md. The guard records an entry in this log file whenever a skill execution count exceeds its configured cap.

If skill caps are missing when checking token usage, the guard applies sensible defaults and writes them to memory/skill-caps.json. This ensures autonomous operations remain constrained without explicit initial configuration.

Use a daily run limit guardrail for agent scheduling when you need to catch expensive scheduled workflows firing too many times, such as a vulnerability scanner running 5 times instead of an expected maximum of 2.