use-a-prompt

Pass

Audited by Gen Agent Trust Hub on Oct 9, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill's primary purpose is to discover, retrieve, and follow external instructions (prompts) via the listPrompts and getPrompt functions. This creates a surface where malicious or unauthorized instructions could be ingested into the agent's context from an external source.
  • Ingestion points: Untrusted content enters the context through the output of the getPrompt capability (SKILL.md).
  • Boundary markers: The skill provides strong instructional boundaries, explicitly stating "Review the prompt with the person who owns the task" and "Follow its approval requirements before executing it." It also includes a warning that prompts are source text and do not grant permission to bypass security constraints.
  • Capability inventory: While specific executable tools are not listed in the frontmatter, the instructions imply that the agent will eventually "execute" or follow the instructions contained within the retrieved prompt.
  • Sanitization: The skill relies on mandatory human review (Step 4) to sanitize and validate the content before any action is taken.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 9, 2026, 07:22 AM