reasoning-controls
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION] (SAFE): The skill defines conceptual slash commands like
/reasoning:adjustand/reasoning:budget. These are configuration instructions for the agent's internal reasoning engine and do not execute arbitrary shell commands or system-level processes. - [EXTERNAL_DOWNLOADS] (SAFE): No external URLs or remote script downloads were found. All references point to local markdown files within the skill's own directory structure.
- [DATA_EXFILTRATION] (SAFE): There are no network operations, credential exposures, or attempts to read sensitive system files. The skill focuses entirely on internal model parameters.
- [PROMPT_INJECTION] (SAFE): The instructions use standard directive language to guide the AI's behavior. It does not contain bypass markers, 'jailbreak' attempts, or instructions to ignore previous safety constraints.
- [DYNAMIC_EXECUTION] (SAFE): The skill contains no code (Python, JavaScript, etc.), only Markdown documentation. There is no risk of runtime code generation or unsafe deserialization.
Audit Metadata