group-tech-design
Pass
Audited by Gen Agent Trust Hub on Oct 25, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill possesses a potential attack surface for indirect prompt injection because it incorporates data from existing external documents into the agent's working context.
- Ingestion points: The
yuque_searchtool (SKILL.md, Step 2) is used to fetch existing documents, which are then used as context for generating new technical designs. - Boundary markers: The instructions do not define specific delimiters or instructions to treat searched content as untrusted data.
- Capability inventory: The skill has the ability to list team repositories (
yuque_list_repos) and write new documents to the remote knowledge base (yuque_create_doc). - Sanitization: A mandatory human-in-the-loop checkpoint is defined in Step 4 ("Review with User"), requiring the user to verify the draft content before any write operations are performed, which significantly mitigates the risk of automated malicious actions.
- [DATA_EXPOSURE]: The skill refers to the use of a 'group Token' for authentication. This is correctly identified as a configuration requirement for the underlying MCP server rather than a request for the user to disclose secrets within the chat prompt, following standard secure configuration practices.
Audit Metadata