Self improving security layer for developers and agents
Security that learns your code.
Prismor is a self-improving security layer for AI agents and modern supply chain defense.
- A patch agent that gets better with use : For both local and PR time
- An immune layer which secures every agent action : Available for local self-hosting and IAM for enterprise
- Compliance artifacts written in real time : Available locally and Enterprise platform
| Repository | Description |
|---|---|
| immunity-agent | Runtime security layer for AI coding agents. Blocks dangerous commands, prevents secret leaks, and enforces runtime policies. |
| product-sbom-vex | Last-mile EU CRA compliance tooling. Generate product-level SBOMs and VEX documents based on CSAF 2.0. |
| security-playbook | Security skill pack for AI coding agents: OWASP code and LLM rules, behavioral guardrails, static analysis guidance. |
| tagdiff | Aggregates changelogs across GitHub releases between two versions. |
PRs are welcome across all repos. For immunity-agent, new detection rules go in warden/default_policy.yaml. Open an issue first if you are unsure where something fits. Run pytest before opening a PR.
- Discord - questions, ideas, and discussion
- X / Twitter - updates and announcements
- prismor.dev - docs and product info
- contact@prismor.dev - everything else
