The Agent Skills Directory

[SAFE]: The skill gathers session context using standard git commands (git diff, git log), which is a benign and secure method for identifying code changes.\n- [SAFE]: File discovery and modification operations are restricted to markdown files (.md) within the project root and the docs/ directory, adhering to the principle of least privilege.\n- [SAFE]: The skill interacts with an internal environment variable (CLAUDE_PLUGIN_ROOT) only to access localized templates for report formatting, which does not involve sensitive data exposure or network activity.\n- [SAFE]: While the skill ingests data from repository files (ingestion points: git output and markdown files), the risk of indirect prompt injection is minimal given the skill's restricted capabilities (file editing and safe bash commands) and its primary purpose of documentation maintenance. Boundary markers and explicit sanitization are absent, but the operations are confined to standard repository management.

up-repo