The Agent Skills Directory

[SAFE]: The skill performs read and write operations exclusively on local files within the memory/ directory and the MEMORY.md file. It manages internal state (heartbeat-state.json, focus-state.json) and daily logs without accessing sensitive system paths or credentials.
[SAFE]: No network activity, external downloads, or remote code execution patterns were detected. The skill does not use tools like curl, wget, or pip.
[SAFE]: The algorithm and rules are consistent with the skill's stated purpose of end-of-day consolidation and do not contain hidden instructions or obfuscated content.
[SAFE]: While the skill processes untrusted input (customer conversation logs) to update long-term memory, the resulting changes are strictly formatting and summarization tasks with no path to arbitrary command execution or privilege escalation.

daily-recap