safety-synthetic-witness-flagger
Pass
Audited by Gen Agent Trust Hub on Jun 24, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill instructions and metadata contain no evidence of malicious intent, obfuscation, or dangerous operations. The logic is entirely focused on providing a forensic checklist for evaluating witness statements.
- [PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted text documents (depositions, affidavits), which is a common vector for indirect prompt injection.
- Ingestion points: Witness statements and deposition transcripts provided to the agent for triage (referenced in SKILL.md).
- Boundary markers: There are no explicit instructions to the agent to disregard instructions that may be embedded within the witness testimony.
- Capability inventory: The skill has no access to sensitive capabilities; it does not utilize file system tools, network requests, or subprocess execution.
- Sanitization: The skill does not implement text sanitization, but given the lack of executable capabilities, the potential impact of an injection is negligible.
Audit Metadata