#evals

Proof Gradient is the agent evolution protocol where every run leaves proof, every proof selects intelligence, and every selected artifact evolves the network.

Evidence, governance, and static reports for agentic runs

Playbook for PMs shipping AI products with PRDs, evals, HITL, launch gates, cost, and observability.

Open-source Agent Skill for Claude Code that makes Claude reason with explicit probabilities, base rates, and falsifiers. Distilled from 51 books; first in 18/18 blind passes vs baseline Claude and GPT-5.5, and benchmarked against 14 rival skills and prompt techniques with full cost telemetry (round 5).

Evidence router & policy engine for coding agents — enforces source choice and proof-of-use through Claude Code hooks, including routes to MCP tools. Local-first; no LLM in the hook loop.

Judgment frameworks for AI agents — Canvas → Govern → Scope → Score. Free Labs skills for Claude Code, Cursor, and eval handoff.

Prove your Agent Skill works before you publish it.

Tests exact agent interventions—skills, prompts, models, tools, and workflows—and publishes what changed, what failed, and what decision follows.

Local-first AgentOps flight recorder and capability firewall for MCP-based coding agents with OpenTelemetry traces.

Product-foundry control plane for AI coding agents: validator gates, buyer scoring, compute allocation, and Demo Day-style investment decisions across startup product candidates. Next step would be in automating outreach to get real PMF validation and iterate from there.

AI/LLM automation reference: RAG, tool calling, human review, evals, observability, safety gates, and local deterministic samples.

Integrity scenarios for agent runtimes, distilled from real fleet incidents — deterministic fault-injection checks against real SDKs

MCP server exposing multivon-eval + pdfhell as agent-callable tools. Drop into Claude Desktop, Cursor, Cline.

Does following a negated instruction come free with the positive one? A reproducible eval — two pre-registered nulls for frontier Claude.

Framework open source para evaluar agentes LLM con function calling / tool use. Scoring por dimensión (tool accuracy, response quality, safety) + CLI, API y dashboard.