manage-supply-chain
Pass
Audited by Gen Agent Trust Hub on Oct 24, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes platform-specific MCP tools, specifically
harness_listandharness_create, to manage organizational resources and define supply chain security policies. These operations are essential for the skill's documented administrative functions. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to process external data, such as dependency inventories and build artifacts, to generate SBOMs and evaluate compliance policies. This creates a surface where external data could attempt to influence agent behavior; however, the skill mitigates this by using established security frameworks (OPA, SLSA) and tools (Trivy, Syft, Cosign).
- Ingestion points: Build tool outputs and dependency lists processed during SBOM generation (Step 3).
- Boundary markers: Not explicitly defined in instructions, but enforced via tool-specific schemas.
- Capability inventory: Resource listing (
harness_list) and policy creation (harness_create). - Sanitization: Dependent on the underlying Harness MCP server and the security tools (Sigstore/Cosign) utilized.
- [SAFE]: References to external security tools like Cosign, Syft, Trivy, and Sigstore are consistent with supply chain security best practices and do not constitute malicious external dependencies or unauthorized remote code execution.
Audit Metadata