Shared contracts for verdict: the critique() + capture interfaces and the Finding / Critique / wire result types.

OpenAPI client for @dimrev4/pixel-verdict-sdk

A marketer's operating mindset as an installable agent skill and CLI: test-size floors, kill rules, honest verdicts on finished tests, which funnel step can fund a verdict, which channel can actually fund a verdict, the lift a volume can read, a queue of

Browser-driven UI/UX verification for coding agents. No commitment, no verdict.

Integration layer: wires resolver, adapters, candidate, checks, diff, and policy into a parity run, and surfaces the verdict (CLI now; GitHub Action surface to follow).

Terminal review surface for coding agents: annotate plans and diffs, return a verdict, send structured feedback back

Programmatic SEO audit engine: 66 rules across 4 categories (integrity, discoverability, citation, data) for SpamBrain risk + AI Overview citability. Verdict ladder + site classifier.

Verdict/Violation contracts, profile-as-code and deterministic judg3d reports.

verdict critique pipeline: the critique() entry point, the grounding / hallucination gate, and the OpenAI-compatible model client.

verdict capture worker: deterministic Chromium capture, DOM geometry, and the code-computed contrast / overflow / touch-target checks.

Shared TypeScript types for the AEG (Agentic Engineering Governance) engine: tranches, tasks, derived status, forge facts, and the check/verdict shapes the other engine packages exchange.

The deterministic agent-integrity gate. One ruleset, evaluated on the actual diff/commands as a verdict, enforced everywhere a change can be made.

Deterministic evaluation of causal rules over BitGraph proof bundles: a pure function from (rule, verified evidence) to a reproducible verdict.

Self-contained HTML comparison page per parity run: reference | candidate | diff side by side with the verdict findings, inlined to one offline .html file (data-URI PNGs + inline CSS/JS, no external assets).

The Brain Layer for AI Trading Agents — one MCP call returns a composite trade verdict (direction, confidence, regime) across major crypto perp venues. On-chain Merkle-verified track record.

Hunter-Seeker Verdict layer for n8n: score, rank, verify, report outcome, drift.

A minimal permission gate for Pi in the style of Claude Code's auto mode

Trufo's policy and verdict engine on top of c2pa-rs. Powers verify.trufo.ai.

Pure verdict algebra, claim matching, hash chain and reconciliation. No I/O.

Infra: the wicked evidence ledger — verdict/run/scenario DomainStore + oracle queries + additive SQLite migrations. Shared by QE writers and gate readers.