Industry · Security operations
Investigate without spreading the blast radius.
Security agents read hostile input all day. Keep each investigation isolated, keep credentials out of the prompt, and decide which tools an agent may call.
How it fits
Built from the primitives you already get.
A workspace per investigation
Logs, samples and notes for each case live in their own filesystem, searchable with globs and full text.
Nothing to exfiltrate
Scanner and SIEM tokens are injected per request and redacted from responses. A prompt-injected agent can't leak a key it never saw.
Policy at the primitive
Tool allowlists and secret grants are enforced by the engine, so a manipulated prompt can't widen what the agent can do.
Full audit trail
Every write and secret use is recorded, so you can reconstruct what an agent did during an incident.
Build
Local-first development.
Build and test on your machine with the open-source engine. No cloud account needed: every primitive runs locally, with the same API as the cloud.
Install the CLI, then run primitives dev. Quickstart
Open source
Built in the open. Contribute freely.
The SDKs, the MCP server and the single-node engine are open source and actively maintained. Submit PRs, report issues, improve docs, or just star the repo to show support.
useprimitives/primitivesFork, star, or open a pull request