scheduled-task-resilience
Pass
Audited by Gen Agent Trust Hub on Oct 13, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill includes shell script templates that use
sleepand the GitHub CLI (gh). These are used to implement robust polling mechanisms for asynchronous operations like CI builds. - [INDIRECT_PROMPT_INJECTION]: The skill involves processing data from external sources, specifically the output of GitHub pull request checks.
- Ingestion points: External data enters the agent context via the output of
gh pr checksinSKILL.md. - Boundary markers: The provided script uses
jqto filter and map external output into a fixed set of status strings (PENDING, FAILING, PASSED), which limits the influence of external data on the agent. - Capability inventory: The skill facilitates shell execution and uses progress storage tools.
- Sanitization: The script explicitly normalizes external JSON output into a controlled set of text tokens using
jq.
Audit Metadata