occ-verify-release
Pass
Audited by Gen Agent Trust Hub on Oct 12, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill involves executing verification scripts, such as
scripts/package-e2e.mjs, to test the functionality of the package release in a local environment. - [EXTERNAL_DOWNLOADS]: The agent is instructed to fetch npm metadata and tarballs for the
@ruvnet/open-claude-codepackage, which is an official resource associated with the skill author. - [INDIRECT_PROMPT_INJECTION]: The skill processes data from external sources, which constitutes an attack surface for indirect prompt injection.
- Ingestion points: npm registry metadata, GitHub Actions logs, and release receipts from the file system.
- Boundary markers: No explicit delimiters are specified for the ingested content.
- Capability inventory: The skill possesses the ability to execute shell scripts and perform network downloads.
- Sanitization: The skill mitigates potential risks by explicitly instructing the agent to disable lifecycle scripts during installation and warning against placing credentials in command arguments or evidence files.
Audit Metadata