CredShields

CredShields on Hugging Face

CredShields blends AI automation with senior-expert auditing to secure smart contracts, protocols, and enterprise applications. We focus on predictive security, developer-first workflows, and standardized risk communication inspired by OWASP. Our flagship is SolidityScan, built to catch logic flaws before mainnet while supporting human-in-the-loop review.

Mission

• Predict and prevent exploits before deployment.
• Standardize risk across ecosystems with OWASP-style clarity.
• Accelerate secure shipping with developer-first tooling and automation.

What We Publish Here

• Models: Solidity-aware LLMs for code understanding, and reasoning.
• Spaces: Lightweight demos for instant scans, diff-based fixes, and rule explainability.

Flagship: SolidityScan

SolidityScan powers fast, predictive scanning for developers. It augments manual reviews with AI-driven findings, risk normalization, and actionable remediation guidance—reducing audit cycle time without compromising depth.

Responsible Use

• These models assist auditors and developers; they do not replace manual review.
• Findings require human validation and protocol-level context before action.

Governance and Contributions

• Use Hugging Face discussions for issues, feature requests, and disclosures.
• We follow semantic versioning and provide changelogs and reproducible evals.
• Contributions are welcome.

Licensing

• Models: permissive licenses (e.g., Apache-2.0) unless otherwise specified in each repo.
• Datasets: open licenses suitable for research and commercial audit workflows.
• Please check each model/dataset card for exact terms.