escalation-flagger
Pass
Audited by Gen Agent Trust Hub on May 15, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill's logic is entirely instructional and operates within the local project context. There are no indications of malicious code, privilege escalation, or persistence mechanisms.
- [PROMPT_INJECTION]: The skill processes untrusted content by quoting contract provisions in drafted messages. While this presents a surface for indirect prompt injection, the risk is mitigated by the 'human-in-the-loop' requirement where a lawyer must review and manually send the drafted output.
- Ingestion points: User-provided problem descriptions, contract quotes, and local files such as 'acme-review-memo.md'.
- Boundary markers: None identified in the provided templates.
- Capability inventory: Limited to text generation and local file reading; no shell execution, network operations, or file-write tools are utilized.
- Sanitization: The skill does not implement specific sanitization or escaping for contract quotes.
- [DATA_EXFILTRATION]: No network-capable commands or unauthorized data transfer patterns were detected. The skill's access to local files is necessary for its stated purpose of reading legal guidelines.
Audit Metadata