dependency-maintenance
Pass
Audited by Gen Agent Trust Hub on Oct 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill involves reading and acting upon repository manifests and scripts, which serves as a potential surface for indirect instructions.\n
- Ingestion points: SKILL.md (Step 1 Discover: 'Read affected manifests and scripts...').\n
- Boundary markers: No explicit delimiters are defined for processing external manifest data.\n
- Capability inventory: SKILL.md ('Run root pnpm install', 'Apply the root changeset routing'), and file write access for manifest updates.\n
- Sanitization: The skill includes a manual verification step, instructing the agent to 'Inspect warnings and the semantic lockfile diff for... lifecycle scripts, and allowBuilds effects'.\n- [SAFE]: The skill represents standard development operations for dependency management and incorporates best practices for verifying automated tool outputs.
Audit Metadata