The Agent Skills Directory

[SAFE]: The skill defines a safety framework for risk classification. It provides heuristics and rules for agents to follow when assessing the risk level of a task based on the files being modified and the nature of the changes.
[SAFE]: No malicious code, prompt injections, or unauthorized network operations were found. The skill references other components as part of its documented integration pattern for high-risk operations.
[SAFE]: The pseudocode provided in the heuristic classifier is intended to guide the agent's reasoning process and does not involve the execution of untrusted code or scripts.
[SAFE]: The skill's override and escalation mechanisms follow security best practices by requiring human intervention for the de-escalation of critical tasks.

risk-classification