Deep learning in Rust, with shape checked tensors and neural networks.
clawdbot-formal-models
This repository contains machine checkable security models for Clawdbot, including regression cases checked with TLC.
About clawdbot-formal-models
This GitHub repository contains machine-checkable security models for Clawdbot, written mainly in TLA+ and checked with the TLC model checker. It is organized as a security regression suite. Green models are expected to pass with no invariant violation. Red models are deliberately buggy variants that are expected to fail and produce a counterexample trace.
The repository includes a quickstart that requires Java 11 or later (Java 21 recommended) and runs individual targets with make. Some checks are conformance-style: they extract tool-group expansions and tool-name aliases from the Clawdbot source into a generated file and check the formal model against those constants. Continuous integration currently runs the TLC targets directly in this repository. A conformance job that checks out the main Clawdbot repository is described as planned, not as existing. The repository also lists deeper modeling of pairing-store concurrency, provider-specific ingress behavior, and routing identity-link semantics as future work.
Key features
- TLA+ specifications checked with the TLC model checker
- Paired green models expected to pass and red models expected to fail
- Counterexample traces produced by deliberately buggy red models
- make targets for running individual models; Java 11+ required
- Conformance script extracting tool-group expansions and aliases from Clawdbot source
- Design notes, threat models, and a security claim inventory in docs
Best for: Engineers working on Clawdbot or reviewing its security claims who want formal, repeatable checks of specific properties.
Good for
- Framed as a practical regression suite, with green and red cases that show what passing and failing look like
- Includes run instructions, a documented structure, and key docs for interpreting results
- Conformance checks tie some model constants to the Clawdbot implementation
Watch out for
- Conformance checks currently cover only tool-group expansions and tool-name aliases
- A conformance CI job and counterexample trace uploads on pull requests are planned, not described as current
- Pairing-store concurrency, provider-specific ingress, and routing identity-link semantics are listed as areas for future modeling
clawdbot-formal-models at a glance
- Type
- Open-source project
- Pricing
- Not confirmed. Check the site
- Open source
- Not stated
- Language
- English
- Last checked
- Oct 6, 2026 · Online when checked
- Listed
- Oct 5, 2026
clawdbot-formal-models alternatives
Tools and sites you could use instead, ranked by how closely they match.
The tensorflow-upstream repository provides a ROCm port of TensorFlow for the project.
Open source
24 alternatives
PromptLayer supports prompt versioning, LLM evaluations, and production monitoring with traces, logs, and regression sets.
Candle provides a minimalist Rust framework for machine learning model development and execution tasks.
Open source
24 alternatives
MLOps framework to package, deploy, monitor and manage thousands of production machine learning models.
Kubeflow provides a machine learning toolkit for deploying workflows on Kubernetes.
Open source
24 alternativesPopular in AI Infrastructure & MLOps
Well-known listings from the same topic.
Guild AIguild.aiFlow AIflow-ai.comsystemprompt.iosystemprompt.ioCoreWeavecoreweave.comKili Technologykili-technology.comAPI docsdocs.x.ai
clawdbot-formal-models FAQ
What is the difference between green and red models?
Green models should pass with no invariant violation. Red models are deliberately buggy variants that should fail and produce a counterexample trace.
What Java version is needed to run the models?
The repository requires Java 11 or later, and Java 21 is recommended.
Where can I find guidance on running the models and reading the results?
The file docs/formal-models.md explains how to run the models, interpret green and red results, and suggested CI approaches.
Is something wrong or out of date?
Tell us and an editor will check it.
Own this website?
Claim this listing
Claim the listing and suggest updates yourself. An editor checks every claim first.