io.github.Olum289/agentlock
Security gateway for AI agents: policy, approval, and audited execution, no secrets shared.
Verdict not yet evaluated for this tool. The hybrid eval runs adversarial cases first; coverage rolls out as the corpus expands. Until a verdict is recorded, an agent should treat this tool as not-yet-cleared and fall back to its own checks. Method: hybrid eval, four-state verdict, honest limits.
Streamable HTTP / SSE endpoint. Add to any MCP client that supports remote servers.
https://app.agentlock.net/api/mcpSecurity tools for AI agents: scan MCP servers, validate HDP delegation chains, audit releases.
Provide AI-powered real-time analysis and intelligence on NPM packages, including security, depend…
Zenable cleans up sloppy AI code and prevents vulnerabilities with deterministic guardrails