The Agent Skills Directory

[SAFE]: The skill's primary purpose is to provide a template and process for documenting 'Lessons Learned' after incidents or mistakes. Its activities are confined to structured data collection and local documentation updates.
[DATA_EXPOSURE]: The skill uses tools like Read, Write, Edit, Glob, and Grep to manage incident logs in log/incidents/ and update project documentation such as MEMORY.md and CLAUDE.md. There is no evidence of sensitive data exposure or exfiltration to external systems.
[REMOTE_CODE_EXECUTION]: No external packages, remote scripts, or network-based execution patterns were detected. The skill operates entirely on the local environment using the agent's built-in file manipulation tools.
[INDIRECT_PROMPT_INJECTION]: While the skill processes user-provided descriptions of incidents (untrusted data) to update documentation or rules (Phase 6), this behavior is the core intended functionality for self-correction. The risk is minimal as it targets structured documentation updates within the local project context.

postmortem