handoff
Pass
Audited by Gen Agent Trust Hub on Sep 6, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external project artifacts which represents a potential attack surface for indirect prompt injection.
- Ingestion points: Reads from .agents/projects/ directory (e.g., CONTEXT.md, LINKS.md).
- Capability inventory: Reads project files and writes summarized handoff files to the filesystem.
- Boundary markers: No specific delimiters mentioned for reading content.
- Sanitization: Explicitly requires redacting API keys, passwords, and personally identifiable information before persistence.
- [SAFE]: The skill implements best practices for secret management by requiring redaction of sensitive credentials.
Audit Metadata