starlight-autoplan
Pass
Audited by Gen Agent Trust Hub on Oct 1, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is primarily a planning framework that incorporates safety by design. It explicitly prohibits the agent from performing destructive or costly actions without human direction and requires verification at every step.
- [INDIRECT_PROMPT_INJECTION]: The skill processes external data (repository evidence and user proposals) to generate plans. While this represents a potential injection surface, the risk is managed through a structured workflow and human gates.
- Ingestion points: User proposals and repository/system evidence processed during the workflow steps in SKILL.md.
- Boundary markers: The skill lacks explicit data delimiters but requires the agent to restate goals and assumptions to ensure context alignment.
- Capability inventory: Includes repository search, testing tools, linting, and browser connector usage for gathering evidence.
- Sanitization: Mitigated by requiring 'human gates' for sensitive operations, 'verification receipts' for promotion, and explicit 'non-goals' to limit scope.
Audit Metadata