ddt
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is instructed to ingest data from external project files to guide its refactoring logic, creating a vulnerability surface where untrusted repository content can influence agent behavior.
- Ingestion points: The skill reads
CLAUDE.md,AGENTS.md, and local module-level documentation files as specified in the "先读项目规范" (Read project specifications first) section ofSKILL.md. - Boundary markers: None specified. The instructions explicitly direct the agent to prioritize the content of these external files over its own default rules.
- Capability inventory: The skill possesses the capability to perform broad file system modifications (refactoring code) and execute validation tools (linting and type checking) as part of its "通用流程" (General Workflow).
- Sanitization: There are no instructions for sanitizing or validating the content of these external configuration files before processing them.
Audit Metadata