gentle-ai
Pass
Audited by Gen Agent Trust Hub on Nov 4, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: The skill includes an 'Identity Rule' that defines the agent's persona as 'el Gentleman,' a senior architect harness. This is a functional steering mechanism for the agent's tone and methodology rather than an attempt to circumvent safety filters or ignore system-level instructions.- [COMMAND_EXECUTION]: The skill references development-related commands such as 'git diff' and automated testing. It explicitly states that 'dangerous-command safety remains independent and authoritative,' indicating that it relies on the underlying platform's security controls for command execution.- [INDIRECT_PROMPT_INJECTION]: The skill incorporates an 'Evidence-budget rule' that limits the volume of external data the agent processes to 10k tokens per batch. This serves as a safety measure to prevent context-stuffing attacks or the ingestion of excessive untrusted data during code reviews and exploration.
Audit Metadata