skill-creator
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to 'Test the skill' by executing realistic test prompts provided by the user (Step 5). This creates a capability surface where the agent processes and acts upon untrusted data as if it were valid instructions.
- Ingestion points: User-provided test prompts (Step 5) and existing repository documentation used for grounding.
- Boundary markers: Absent; the skill does not suggest delimiters or 'ignore' instructions when executing test prompts.
- Capability inventory: File system read/write access (for creating/updating SKILL.md files) and the ability to process arbitrary prompts during the testing loop.
- Sanitization: No sanitization, escaping, or validation of user-provided test prompts is suggested before the agent executes them.
Audit Metadata