printing-press-output-review
Pass
Audited by Gen Agent Trust Hub on May 11, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes the
printing-pressscorecard utility and provides instructions for a sub-agent to invoke the CLI binary at$CLI_DIR/<cli-name>-pp-clito verify findings. This is legitimate behavior for a diagnostic review skill. - [PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface as it analyzes untrusted output from external CLI commands.
- Ingestion points: The skill reads command output samples from
/tmp/output-review-livecheck.json(specifically theoutput_samplefield). - Boundary markers: The prompt contract for the Reviewer Agent does not include explicit delimiters or instructions to treat the sampled output as untrusted data.
- Capability inventory: The Reviewer Agent has access to the
Agenttool and is explicitly authorized to execute the CLI binary. - Sanitization: No sanitization or escaping is performed on the sampled command output before it is processed by the agent.
Audit Metadata