The Agent Skills Directory

[SAFE]: The skill accesses local telemetry logs (.claude/logs/subagent-spawns.jsonl and ~/.claude/analytics/agent-usage.jsonl) to analyze agent usage. This access is consistent with its stated purpose of auditing agent activation.\n- [SAFE]: The core logic is implemented in a local Node.js script (scripts/collect-activation-audit.mjs) which uses standard libraries (fs, path) to process data deterministically without external dependencies or remote code execution.\n- [SAFE]: The skill includes validation logic that matches telemetry data against a local catalog of agent files on the disk, preventing the processing of arbitrary or unverified agent names from the logs.\n- [SAFE]: No network operations, credential exfiltration patterns, or obfuscation techniques were detected in the skill instructions or scripts.

audit-activation