Keyhole – Share secrets with your AI agents
Agents get keychain references instead of values — no secrets in chat transcripts.

Explains runtime/secret but it's a blog post, not a tool or library you can use today.
Go developers concerned about secret management and security
Agents get keychain references instead of values — no secrets in chat transcripts.
Memory protection crate with guard pages and zeroization, like memsec but with secrecy-style ergonomics.
The plugin-proxy split is smart: credentials live in a backend (Keychain/1Password/Vault/etc.) and a separate proxy injects auth headers over a UDS so the agent process never handles raw keys. It autosurveys plugin configs and channels to migrate plaintext secrets and even ships a Docker image and CLI for local setups — very practical for anyone already on OpenClaw, though it’s narrowly focused and adds an extra trusted component that deserves an audit.
Network-layer proxy that injects secrets per request so agents never actually hold the keys.
Secret substitution proxy keeps credentials out of agent context—clever security model.
AI agents get credentials without ever seeing them—SQL prepared statements for secrets.