consultant
Pass
Audited by Gen Agent Trust Hub on Aug 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONPERSISTENCE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external sources identified as "supplied facts" and "target territory." This design allows for the possibility of indirect prompt injection if those files contain malicious instructions.
- Ingestion points: SKILL.md (Step 2: Inspecting the supplied facts and target territory).
- Boundary markers: The skill lacks explicit markers or instructions to the agent to ignore potentially embedded prompts in these sources.
- Capability inventory: The skill can write to the local file system to save state (Step 4) and draft plans (Step 5).
- Sanitization: No sanitization or escaping of external content is specified.
- [PERSISTENCE]: The skill maintains context across sessions by writing "durable state" to
~/.consultant/and copying artifacts into local subdirectories. This is a legitimate functional requirement for a consultation tool to maintain a consistent state across turns.
Audit Metadata