Glossary
Configuration Fingerprint
A configuration fingerprint is a SHA-256 cryptographic hash computed from an AI agent's complete configuration -- its model, prompt template, tools, memory stack, and RAG sources -- that uniquely identifies a specific agent build.
What is Configuration Fingerprint?
Just as a human fingerprint uniquely identifies a person, a configuration fingerprint uniquely identifies a specific version of an AI agent. It is computed by taking the SHA-256 hash of the agent's full configuration: the model identifier, the system prompt, the list of integrated tools, the memory architecture, and the RAG data sources. Any change to any of these components produces a different fingerprint.
Configuration fingerprinting solves a fundamental problem in agent trust: how do you know that the agent you evaluated yesterday is the same agent you are interacting with today? Without fingerprinting, an operator could quietly swap an agent's model, change its prompt, or alter its data sources, and counterparties would have no way to detect the change. The fingerprint makes configuration changes detectable and accountable.
The fingerprint is computed using SHA-256, a cryptographic hash function that is deterministic (the same input always produces the same output), collision-resistant (it is computationally infeasible for two different configurations to produce the same fingerprint), and one-way (the fingerprint cannot be reverse-engineered to reveal the configuration). This means operators can prove their configuration is unchanged without revealing proprietary details.
Example
An agent's configuration fingerprint is computed from: model="claude-opus-4", prompt_hash="a3f8b2...", tools=["stripe-api","zendesk-api","internal-crm"], memory="conversation-buffer-30d", rag_sources=["product-docs-v3","support-kb-v7"]. The SHA-256 hash of this concatenated data produces fingerprint "7c3e9f...". When the operator updates the product docs to v4, the fingerprint changes to "2b8a1d...", triggering an 8% score decay for a RAG source change.
How Signet addresses this
Configuration fingerprinting is a core mechanism of the Signet platform. Every registered agent's configuration is fingerprinted at registration and re-fingerprinted whenever changes are detected. Signet tracks the full fingerprint history, enabling counterparties to see how frequently an agent's configuration changes and what types of changes have occurred. The fingerprint is the anchor that connects an agent's score to its actual current configuration.
Build trust into your agents
Register your agents with Signet to receive a permanent identity and trust score.