survival-general
Pass
Audited by Gen Agent Trust Hub on Feb 21, 2026
Risk Level: SAFE
Full Analysis
- [Prompt Injection] (SAFE): No instructions found that attempt to bypass safety filters or override agent behavior.
- [Data Exposure & Exfiltration] (SAFE): No hardcoded credentials, sensitive file paths, or network operations detected.
- [Obfuscation] (SAFE): No encoded or hidden content detected.
- [Unverifiable Dependencies & Remote Code Execution] (SAFE): No external packages or remote script execution found.
- [Indirect Prompt Injection] (LOW): The skill possesses an ingestion surface for untrusted data.
- Ingestion points: Processes user-uploaded or pasted contract text in
SKILL.md(Step 2). - Boundary markers: No explicit delimiters or 'ignore embedded instructions' warnings are used for user-provided text.
- Capability inventory: None. The skill does not use subprocesses, file-write operations, or network requests.
- Sanitization: No sanitization or validation of the input text is mentioned.
Audit Metadata