empirical-validation
Pass
Audited by Gen Agent Trust Hub on Oct 3, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill requires the agent to verify work by executing commands and interacting with UIs, which involves processing output from potentially untrusted external sources like web pages or API responses. This creates a surface where malicious instructions hidden in that data could influence the agent.
- Ingestion points: Tool outputs from
run_commandandbrowser_subagent(SKILL.md). - Boundary markers: The skill does not define specific delimiters or instructions to ignore embedded commands within the verification evidence.
- Capability inventory: The skill employs
run_command(subprocess execution) andbrowser_subagent(web interaction). - Sanitization: No explicit sanitization or validation steps are prescribed for the ingested evidence before it is evaluated by the agent.
Audit Metadata