ml-pipeline
Pass
Audited by Gen Agent Trust Hub on Oct 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill analyzes existing codebases and repository instructions, creating a surface for indirect prompt injection if those files contain malicious instructions.
- Ingestion points: Inspects available files, inputs, and repository instructions to plan implementations (SKILL.md).
- Boundary markers: While instructions emphasize verification and bounded outcomes, they do not explicitly define delimiters to isolate untrusted data.
- Capability inventory: The skill is capable of implementing code changes, creating artifacts, and running verification tests.
- Sanitization: The instructions do not specify sanitization for ingested content, relying instead on verification of implementation outputs.
- [SAFE]: The skill utilizes vendor-specific tools (sandbase_*) for documentation and inference, following secure interaction patterns such as secret isolation, data minimization, and explicit user consent for paid operations.
Audit Metadata