Ecosystem

SkillFortify

Formal verification for every agent framework. One command. Mathematical proof.

Open Source PyPI

Securing the Skills That Contracts Govern

agentAssert defines what agents must do. SkillFortify verifies that the skills agents use are safe to execute. Together, they provide both behavioral governance and supply chain security.

Capabilities

Formal verification that scales across the agent ecosystem.

22 Framework Profiles

Pre-built verification profiles for Claude Code, Cursor, VS Code, Windsurf, and 18 more agent frameworks. Each profile encodes the framework's permission model, skill format, and execution semantics.

Formal Verification

Not heuristics. Not pattern matching. Formal analysis of skill definitions against security properties. Every verification result is a mathematical proof of safety or a concrete counterexample.

SBOM Generation

Auto-generate CycloneDX Software Bill of Materials for every skill in your agent's toolchain. Full dependency tracking, license enumeration, and vulnerability cross-referencing.

Zero False Positives

F1 score of 96.95% across comprehensive benchmarks. When SkillFortify says a skill is safe, it is safe. When it flags a risk, the risk is real. No alert fatigue. No false alarms.

22 Frameworks
675 Tests
96.95% F1 Score
~2.5ms Per Skill

Explore SkillFortify

Open source skill verification for the agent ecosystem.