verify-with-controls
Pass
Audited by Gen Agent Trust Hub on Oct 11, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill includes several bash command examples and templates (using
git,gh,awk,grep) designed to robustly verify project state and tool outputs. These are provided as defensive coding patterns to ensure accuracy in agent reporting. - [INDIRECT_PROMPT_INJECTION]: The skill defines how to handle and verify data ingested from tool outputs (like git diffs or GitHub CLI checks). It specifically addresses the risk of "blind instruments" or "swallowed errors" by requiring evidence-bearing probes and second opinions, which serves as a defensive measure against misleading data encountered during runtime.
Audit Metadata