gentle-ai-chained-pr
Pass
Audited by Gen Agent Trust Hub on Oct 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses the
gh(GitHub CLI) tool to interact with repositories for viewing and creating Pull Requests.\n - Evidence: The file
references/chaining-details.mdcontains commands such asgh pr view <PR_NUMBER>andgh pr create --repo "$TARGET".\n - Mitigation: The instructions explicitly require human authorization for the destination host, repository, operation, and credentials before any commands are executed.\n- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external Pull Requests, creating a surface where malicious text in a PR could attempt to override the agent's instructions.\n
- Ingestion points: Data is read from Pull Requests via
gh pr viewinreferences/chaining-details.md.\n - Boundary markers: The skill uses a structured "Chain Context" template for its output, though no specific delimiters are mentioned for the ingested data.\n
- Capability inventory: The skill has the capability to create new branches and PRs using the GitHub CLI.\n
- Sanitization: No specific sanitization or filtering of the ingested PR metadata is mentioned in the instructions.
Audit Metadata