agent-acceptance-gate
Pass
Audited by Gen Agent Trust Hub on Oct 12, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses deterministic commands to verify success criteria (e.g.,
gh apifor GitHub data verification). These commands are scoped to project-specific verification tasks and are declarative in nature. - [REMOTE_CODE_EXECUTION]: The skill mentions external fact-checking via URLs (e.g.,
openai.com,arxiv.org), but this is restricted to verification and documentation purposes. There is no pattern of downloading and executing arbitrary scripts or binaries. - [DATA_EXFILTRATION]: While the skill reads repository diffs and logs, it lacks any mechanisms to send this data to unauthorized external endpoints. Network activity is limited to official APIs (GitHub) for metadata verification.
- [PROMPT_INJECTION]: The skill contains internal safety logic to prevent agents from bypassing constraints (e.g., 'Never claim PASS' if a preset fails). It explicitly prohibits agents from muting failures or overriding policy decisions.
- [DYNAMIC_EXECUTION]: The presets (YAML files) define structural and pattern-based checks (regex, file existence) rather than dynamic code generation or execution.
Audit Metadata