secondbrain-freshness
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-controlled data (record titles and IDs) from various project entities and interpolates them into a Markdown report. This creates a potential surface where malicious instructions hidden in a file's metadata could be parsed by the agent during report generation.\n
- Ingestion points: The skill reads records from the local file system (ADRs, Tasks, Notes, Discussions) to extract titles and ages in Step 2.\n
- Boundary markers: There are no explicit delimiters or instructions for the agent to ignore instructions that might be embedded within the record fields during analysis.\n
- Capability inventory: The skill is capable of reading project files and updating the
.claude/data/config.yamlfile.\n - Sanitization: The workflow does not specify a sanitization or validation step for item content before it is displayed in the summary and detailed tables.
Audit Metadata