factory-recover
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill identifies interrupted tasks by reading historical conversation data, including original user requests and subsequent messages.
- Ingestion points: Analyzes messages and requests from past runs (SKILL.md).
- Boundary markers: Explicitly instructs to resume only with original authorization and current user instructions, and prevents restoring canceled work.
- Capability inventory: Involves potential file writes, replies, approvals, merges, and deployments mentioned in Step 4 of the recovery process.
- Sanitization: Relies on host-level authorization checks and existing instruction boundaries.
- [EXTERNAL_DOWNLOADS]: The instructions reference configuration guidelines and documentation hosted on the official GitHub repository for the tool (github.com/BuilderIO/skills).
Audit Metadata