lode-weekly-outline
Pass
Audited by Gen Agent Trust Hub on May 12, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill performs read and write operations within a local directory context (referred to as the 'vault'). It manages project-specific reporting by reading 'projects.json' and 'raw' entry files.
- [COMMAND_EXECUTION]: The skill invokes 'git log' to fetch commit history for the requested time range. This command execution is localized to the repository path and used exclusively for data retrieval.
- [DATA_EXFILTRATION]: There are no network-related commands (e.g., curl, wget) or external API calls. The data processing cycle (Scope → Analysis → Stitching) remains entirely local to the user's environment.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from git commit messages and raw change entries. While the subagent prompt in 'subagent-prompt.md' lacks explicit structural delimiters (like XML tags or specific boundary markers) to isolate this data, the risk is limited to the content of the generated Markdown report. The skill does not possess capabilities that would allow for system-level compromise via this vector.
- [PROMPT_INJECTION]: The instructions in 'SKILL.md' are focused on operational logic and do not contain attempts to bypass safety filters or override agent constraints.
Audit Metadata