by joshuaisaact · AI 工具 · ★ 6
PetriFlow Provably safe AI agents. Rules your agent cannot break. Any framework. One safety layer. PetriFlow compiles declarative safety rules into Petri nets that gate every tool call. Each rule is verified exhaustively before your agent starts. If the verifier says "safe," it means safe in every possible execution, not just the ones you tested. Quick start Run an example agent (requires an API key for the model provider): Safety layer packages The core product: a framework-agnostic tool gating system with adapter packages for specific agent runtimes. Vercel AI SDK adapter. Wraps tool meth
| Stars | 6 |
| Forks | 0 |
| Language | TypeScript |
| Category | AI 工具 |
| License | MIT |
| Quality Score | 31.75/100 |
| Open Issues | 1 |
| Last Updated | 2026-03-21 |
| Created | 2026-02-08 |
| Platforms | node |
| Est. Tokens | ~89k |
Explore other popular ai 工具 tools:
petri-flow is Petri net safety layer for AI agents. Compile declarative rules into formally verified tool gates. Any framework.. It is categorized as a AI 工具 with 6 GitHub stars.
petri-flow is primarily written in TypeScript.
You can find installation instructions and usage details in the petri-flow GitHub repository at github.com/joshuaisaact/petri-flow. The project has 6 stars and 0 forks, indicating an active community.
petri-flow is released under the MIT license, making it free to use and modify according to the license terms.