autopilot-spec
Pass
Audited by Gen Agent Trust Hub on Oct 10, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and act upon data from external sources that could contain malicious instructions.
- Ingestion points: In
SKILL.md(Step 1), the agent is instructed to read "repository guidance, the spec, associated tickets, acceptance criteria, and tracker/forge configuration." - Boundary markers: The instructions do not specify any delimiters or safety prompts to distinguish between data and instructions within these external sources.
- Capability inventory: The skill possesses the ability to implement code, perform repairs, commit changes, push to branches, and execute merges (
SKILL.mdSteps 2, 4, and 5). - Sanitization: There is no mention of filtering, escaping, or validating the content retrieved from tickets or specifications before it influences the agent's actions.
- [COMMAND_EXECUTION]: The skill automates several high-impact Git and forge operations.
- The workflow explicitly authorizes the agent to perform commits, pushes, and automatic merges once specific criteria are met, which involves executing shell-level or API-level commands against a repository.
Audit Metadata