archive-to-vault
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes data from external files in the KNOWLEDGE area or direct user input to generate new vault files. This creates a surface for indirect prompt injection where malicious instructions embedded in the source data could influence the agent's behavior during the conversion or writing process.
- Ingestion points: Reads from
~/.claude/memory/KNOWLEDGE/{type}/{slug}.mdand accepts direct entity data. - Boundary markers: Includes a mandatory user confirmation step before file creation.
- Capability inventory: Uses
Write,Edit, andBashtools. - Sanitization: Employs structured mapping of specific sections (Thesis, Background, etc.) to template fields.
- [DATA_EXFILTRATION]: The skill explicitly accesses sensitive internal agent storage paths located in
~/.claude/memory/. Accessing these paths is necessary for the stated purpose of knowledge management but constitutes access to the agent's internal state.
Audit Metadata