fable-handoff
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFENO_CODEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill defines a process for context compaction and handoff generation without using executable code or external resources. No malicious patterns such as prompt injection, obfuscation, or persistence were detected.
- [NO_CODE]: The skill package contains no scripts or executables. All logic is provided via natural language instructions and structured templates.
- [INDIRECT_PROMPT_INJECTION]: The skill has a data ingestion surface through
session_context. 1. Ingestion points: session context history (SKILL.md). 2. Boundary markers: No explicit input delimiters are defined for the session history. 3. Capability inventory: The skill specifies no tool access or scripting capabilities (skill.package.json). 4. Sanitization: Stage 8 of the protocol explicitly requires sanitizing secrets and referencing secure auth instead of copying values. The lack of capabilities renders this surface safe.
Audit Metadata