wait-what
Pass
Audited by Gen Agent Trust Hub on Sep 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill body instructs the agent to process and use content from an external file (
CONTEXT.md). This creates an ingestion point for potentially untrusted data that could influence the agent's behavior if the file is compromised. - Ingestion points: Mention of
CONTEXT.mdin the skill body for interpolation into the prompt. - Boundary markers: The instructions lack delimiters or explicit boundary markers to separate the agent's instructions from the potentially untrusted content of the context file.
- Capability inventory: The skill is restricted; it contains no command execution or network access, and the frontmatter explicitly sets
disable-model-invocation: trueto prevent tool use. - Sanitization: No sanitization, filtering, or validation steps are defined to process the data from the context file before it is provided to the agent.
Audit Metadata