ClawVault
ClawVault is a security layer for OpenClaw-style AI agents that logs and can block what the agent touches, letting a user lock private files, API keys and credentials into a monitored vault and get an instant-messaging alert whenever the agent reaches for them.
ClawVault is a security vault for AI agents built around atomic, per-capability control. Users load private files and set up secure storage, then define which agents, skills, credentials and files count as protected; whenever one of those assets is touched, a notification is sent over instant messaging describing who touched what.
Fine-grained policies, called atomic capabilities, cover agent interaction and invocation, model routing and quota limits, security detection such as sensitive-information recognition, credential detection and prompt-injection protection, and file-access permissions. Event collection combines an API gateway with file-side monitoring, and users can also describe a new detection scenario in natural language and have the system generate and enforce the corresponding policy rule.