captp
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFE
Full Analysis
- [Prompt Injection] (SAFE): No override markers or jailbreak attempts detected.
- [Data Exposure & Exfiltration] (SAFE): No hardcoded credentials or sensitive file paths identified. Network layers mentioned (Tor, Tailscale) are architectural and not active exfiltration.
- [Obfuscation] (SAFE): No Base64, zero-width characters, or homoglyphs found.
- [Remote Code Execution] (SAFE): No patterns for piped script execution or untrusted downloads detected.
- [Indirect Prompt Injection] (LOW): The skill handles distributed message processing. 1. Ingestion points: Message types like op:deliver in SKILL.md. 2. Boundary markers: Protocol enforced via Vats and unforgeable capabilities. 3. Capability inventory: Actor invocation via send commands. 4. Sanitization: Security model based on Principle of Least Authority (POLA).
- [Command Execution] (SAFE): Just commands listed are for protocol management; no auto-executing or dangerous command sequences were identified.
Audit Metadata