Phinq

Phinq provides a critical runtime governance layer for autonomous agents, ensuring operational safety and accountability. Key features include: * Intercepts all agent tool calls * Classifies actions by risk level * Allows safe actions to proceed * Holds irreversible actions for human review * Records all decisions in a tamper-evident audit log This open-source solution acts as a crucial checkpoint, preventing unintended or damaging operations. When an agent proposes a tool call, Phinq evaluates it based on tool name, arguments, and session context. Harmless actions are executed immediately, while those identified as risky are held for explicit human approval through a configured channel. This pause-and-review mechanism ensures that sensitive actions like data deletion, credential access, financial transactions, or external communications cannot proceed without explicit consent. Every decision made—whether an action is passed, held, approved, or denied—is meticulously recorded in a hash-chained audit log. This provides a transparent and verifiable history of all agent activities, crucial for compliance, debugging, and maintaining trust in autonomous systems. Should a risky action remain unanswered, it defaults to a denied status, effectively failing closed to prevent unauthorized execution. Phinq is an essential tool for development teams, security professionals, and compliance officers working with agent-based systems. It's ideal for maintaining control over autonomous processes, mitigating operational risks, and establishing a clear, auditable trail for all agent interactions within an application or enterprise environment.