grooming-backlogs

Pass

Audited by Gen Agent Trust Hub on Oct 30, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from GitHub issues (titles, bodies, and comments) which could potentially contain malicious instructions designed to influence the agent's behavior.\n
  • Ingestion points: Data is read from GitHub Projects and repositories using the gh CLI.\n
  • Boundary markers: The skill includes specific instructions to treat issue data as untrusted and references external safety guidelines such as shared/external-data.md.\n
  • Capability inventory: The skill has the ability to read from and write to GitHub Projects, create local directories, and execute shell commands.\n
  • Sanitization: The skill relies on a mandatory human-approval process for every planned mutation to prevent automated exploitation of injected prompts.\n- [EXTERNAL_DOWNLOADS]: The skill relies on external logic and safety rules hosted on the author's GitHub repository. The instructions direct the agent to read these external files (e.g., hard-rules.md, human-control.md) from https://github.com/bostonaholic/skills/tree/main/skills/engineering/grooming-backlogs.\n- [COMMAND_EXECUTION]: The skill uses the GitHub CLI (gh) and shell commands to perform its tasks. Commands include gh repo view, gh project list, gh project field-list, and mktemp. The gh auth refresh command is also used to ensure the necessary permissions for project management.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 30, 2026, 02:37 AM