discussing
Pass
Audited by Gen Agent Trust Hub on Aug 26, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill directs the agent to autonomously fetch facts from the local environment, including the filesystem and available tools, to resolve its own prerequisite questions. This creates an indirect prompt injection surface where the agent could be influenced by malicious instructions or data found within the files it reads during its context-gathering phase.
- Ingestion points: Data is ingested from the local filesystem and external tool outputs as part of the 'frontier' recomputation logic in SKILL.md.
- Boundary markers: The instructions lack explicit boundary markers or delimiters to separate ingested environmental data from the agent's core instructions.
- Capability inventory: The skill encourages active use of filesystem and environment tools via sub-agents to gather facts.
- Sanitization: There is no requirement for the agent to sanitize, escape, or validate data retrieved from the environment before incorporating it into the discussion context.
Audit Metadata