research
Pass
Audited by Gen Agent Trust Hub on Jun 18, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [SAFE]: The skill is entirely composed of markdown instructions and does not include any scripts, executables, or dynamic code execution.
- [EXTERNAL_DOWNLOADS]: The skill instructs the agent to gather data using web search and documentation tools. It encourages the use of primary sources such as official documentation, repositories, and RFCs, which are standard technical resources.
- [PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection because it processes untrusted data from the web. It attempts to mitigate this through a distillation step that converts raw content into single-line findings, reducing the impact of embedded prose.
- Ingestion points: External web search and primary documentation (SKILL.md).
- Boundary markers: Not explicitly defined for external content.
- Capability inventory: Web search tools, documentation access tools, and data passing to a secondary skill.
- Sanitization: Summarization process (distillation) effectively strips extraneous formatting and prose.
Audit Metadata