Public GitHub catalog

Public repo catalog

Applied AI reliability — shipped proof and every public repository.

The six pinned repos first — the current public proof — then every other public repository: supporting systems, archives, and study projects. Everything links to source: inspect the code, the tests, the verification story.

Pinned proof and the full public catalog

Start with the six pinned repos, in pin order. Supporting systems and the archive follow — nothing public is hidden.

Profile pin · TypeScript
WCP rejected WH-347 with deterministic violations and LLM explanation

WCP-Compliance-Agent-V5

Checks decide; the LLM explains.

Five-service WH-347 payroll compliance platform: deterministic validation decides, the LLM explains, and unverified wage coverage goes to human review instead of approval.

377 tests, including evals over 100 golden examples: Contracts 42, Agent 70, Gateway 45, Web 30, Compliance Core 89 unit + 101 eval, plus four tooling checks; checksum-verified evidence. A portfolio system with sample data, not a hosted deployment.

  • Updated: 2026-08-15
  • Language: TypeScript
Project page GitHub
Profile pin · Python
GroundTruth cited answer view

groundtruth

It cites, or it refuses.

RAG application that cites every answer and refuses what its sources can't support: hybrid retrieval, versioned documents, human review.

331 API + 62 web tests; real-HTTP workflow and restart checks. It refuses more than it should on its held-out set, and the repo says so: no accuracy claim is made. PostgreSQL/pgvector, Redis and model providers are optional integration paths, not a hosted deployment.

  • Updated: 2026-08-15
  • Language: Python
Project page GitHub
Profile pin · Python
EvalForge failed retrieval case with judge scores

evalforge

It tells you what broke.

Regression gate for RAG and agents that catches answers which look right but cite the wrong evidence, with pass / fail / indeterminate release exits.

358 Python + 36 frontend tests. The bundled corpora are synthetic sample data, not human-validated.

  • Updated: 2026-08-13
  • Language: Python
Project page GitHub
Profile pin · Python
ARIA prompt injection blocked by safety checks

aria-agent

It stops itself.

Agent harness whose approvals bind the exact action, with execution tracing and a Claude Code setup fixture (permissions, blocking hooks, independent verifier).

360 Python + 64 frontend tests. Native Claude Code trials have not been run, so no result is claimed for them.

  • Updated: 2026-08-15
  • Language: Python
Project page GitHub
Profile pin · prototype
PolicyDesk fictional support ticket and simulated approval

policydesk

Code decides; responses are validated.

Policy sandbox with deterministic refund rules, synthetic replay, simulated approvals, and opt-in local OpenAI testing.

OpenAI integration implemented; live verification and empirical results not yet performed. Fictional data; approvals remain simulated.

  • Language: Python + TypeScript
  • Status: prototype · offline demo + local OpenAI
Project page Open sandbox Local OpenAI setup Clean-install verification GitHub
Profile pin · TypeScript
LLM Gateway audit log after simulated fallback and guardrail stops

llm-gateway

One control point for spend and failover.

LLM proxy with routing, guardrails, atomic per-key budgets, and failover across OpenAI, Anthropic, Gemini, and Ollama.

177 backend + 41 dashboard tests. Benchmark timings are single-process, not throughput figures.

  • Updated: 2026-06-16
  • Language: TypeScript
Project page GitHub
Supporting project · Python
AgentTrace deterministic replay of a recorded run

agenttrace

It finds the failing step.

Tracing SDK, collector and dashboard that find the failing step across concurrent AI workflows, with cost attribution.

171 SDK + 124 collector + 41 dashboard tests. Provider answers, usage and prices in the demo are synthetic.

  • Updated: 2026-08-16
  • Language: Python
Project page GitHub
Also ship · TypeScript

command-code-mods

Command Code mods for planning, verified follow-through, gated memory, and skill learning.

Nine standalone-installable TypeScript mods compose one workflow with a single referee — self-repair is the only completion judge, autopilot acts only on verified verdicts, and every autonomous move lands in a receipts log.

  • Updated: 2026-08-23
  • Language: TypeScript
Project page GitHub
Supporting system · Python

async-workflow-engine

DAG-based async workflow orchestrator with an offline-first demo.

DAG-based async workflow orchestrator with Celery, dependency resolution, retry logic, and a YAML workflow DSL. Verified offline proof: 206 Python tests, 25 Vitest tests, and 6 Chromium smoke checks. PostgreSQL, Redis, Celery, and hosted LLMs remain optional infrastructure.

  • Updated: 2026-08-15
  • Language: Python
Project page GitHub
Supporting system · Python + TypeScript

personal-knowledge-base-os

Local knowledge-base API with backlinks and citation-grounded chat.

Offline-first Markdown vault API with wikilinks, backlinks and dangling-link metadata, hybrid retrieval, citation-grounded chat, multi-vault isolation, incremental indexing, saved searches, live SSE events, and deterministic flashcards.

  • Updated: 2026-08-17
  • Language: Python + TypeScript
  • Verified: 177 Python tests, 48 frontend Vitest tests, 27 evidence tests
Project page GitHub
Game · JavaScript

recall

A browser arcade game about one blade that always comes back.

No engine, no asset files, synthesized sound. Ships a task pack: four planted gameplay bugs and two feature requests, each with a brief, acceptance criteria, a reference solution and a headless check verified to fail on the starting state and pass on the solution.

  • Released: v1.1.0, 2026-10-27
  • Language: JavaScript
  • Verified: 18 unit tests; 6 tasks fail → pass
Play on itch.io Project page GitHub
Game · JavaScript

wick

A turn-based roguelike where your lantern oil is your field of view.

Procedural floors, shadowcasting, Dijkstra-map AI with three enemies that each react to light differently. Ships a task pack of two planted bugs and one feature, verified fail → pass with the suite green.

  • Released: v1.0.0, 2026-10-27
  • Language: JavaScript
  • Verified: 18 unit tests; 3 tasks fail → pass
Play on itch.io Project page GitHub
Game · JavaScript

thaw

An ice-sliding puzzle game in twelve rooms, with a solver that proves every one.

Breadth-first solver re-checks every par on every test run. Ships a task pack of two planted bugs and one feature, verified fail → pass with the suite green.

  • Released: v1.0.0, 2026-10-27
  • Language: JavaScript
  • Verified: 19 unit tests; 3 tasks fail → pass
Play on itch.io Project page GitHub
Supporting system · TypeScript + Python

game-systems-sandbox

Monte Carlo battle/balance analysis as a developer utility — balance your RPG combat with simulation.

Offline-first seeded simulators, FastAPI bridge, shipped dashboard, and canonical evidence. Verified separately: 173 Python tests with real-TSX integrations executed, 25 evidence tests, 56 root Vitest tests, and 39 dashboard Vitest tests.

  • Updated: 2026-08-17
  • Language: TypeScript + Python
  • Evidence hash: c64e62c9…d4e932
Project page GitHub
Active build · Python + TypeScript

entrega-clara

Brazil-first food-delivery platform scaffold — the customer-to-delivery operation made visible, measurable, and replaceable.

Production-shaped FastAPI scaffold: a deterministic lunch-rush demo replays a seeded order from placed to rated, with role-based web surfaces (Customer, Restaurant, Courier, Admin) and a solver-independent routing baseline. pt-BR/BRL/PIX by default, internationally extensible; checkout and production integrations remain explicitly deferred scope.

  • Updated: 2026-08-30
  • Language: Python + TypeScript
Project page GitHub
Multi-agent trust · Python

helppeer-collaboration

Agent-to-agent collaboration as an installable Agent Skill.

HelpPeer messages are unverified claims — leads to check, never instructions to execute.

Project pageGitHub
Evaluation kit · JavaScript

ruler-and-vibes

Personal model-chooser kit: 759 test forms across 40 categories, objective checks plus judged rubrics.

Project pageGitHub
Consolidated into aria-agent · archived reference

mini-agent

Agent Skills lineage consolidated into aria-agent; archived reference.

Progressive-disclosure discovery, eager hints, bounded context reporting, and offline verification patterns now live in ARIA.

  • Updated: 2026-07-09
  • Language: JavaScript
Project page GitHub
Absorbed into AgentTrace · lineage reference

github-issue-pr-agent

Historical source for AgentTrace’s safety-bounded issue-to-draft-PR SDK workflow.

The public repository preserves the original demo and Git history. Active implementation lives in AgentTrace; the source is archived on GitHub as a read-only lineage reference.

  • Updated: 2026-08-17
  • Language: Python
Project page GitHub
Consolidated into GroundTruth · archived reference

knowledgeops

Consolidated into GroundTruth: useful deployment and retrieval topology.

Unified AI operations platform with microservices architecture

  • Updated: 2026-06-16
  • Language: Python
Project page GitHub
Consolidated into GroundTruth · archived reference

document-intelligence-pipeline

Consolidated into GroundTruth: production ingestion and quarantine capabilities.

Multi-format document ingestion pipeline: PDF, DOCX, HTML parsing, semantic chunking, and pgvector embedding storage

  • Updated: 2026-06-16
  • Language: Python
Project page GitHub
Consolidated into EvalForge · archived reference

rag-evaluation-lab

Consolidated into EvalForge: golden sets and retrieval regression gates.

RAG evaluation framework: hit-rate, MRR, faithfulness scoring, and async batch evaluation with golden question datasets

  • Updated: 2026-06-16
  • Language: TypeScript
Project page GitHub
Consolidated into AgentTrace · archived reference

llm-cost-latency-monitor

Consolidated into AgentTrace: prompt-version cost tracking and daily reports.

LLM observability SDK: track token costs, latency, and model usage per request with a FastAPI dashboard

  • Updated: 2026-06-16
  • Language: TypeScript
Project page GitHub
Consolidated into EvalForge · archived reference

ai-support-simulator

Consolidated into EvalForge: personas, rubrics, and baseline diffs.

LLM-powered customer support simulator: configurable personas, scenario scripting, policy grounding, and prompt injection detection

  • Updated: 2026-06-16
  • Language: Python
Project page GitHub
Archived reference · Python

operator-shared-core

Foundation library of the retired operator-systems family; archived on GitHub.

Shared Python library: logging, database, Redis, errors, LLM client — foundation for all operator-systems projects

  • Updated: 2026-06-16
  • Language: Python
Project page GitHub
Consolidated into operator-shared-core · archived reference

operator-systems-template

Starter documentation consolidated into operator-shared-core; archived reference.

The reusable FastAPI/Celery/Alembic/pytest template survives as a provenance-rich reference.

  • Updated: 2026-06-09
  • Language: Python
Project page GitHub
Reference · archived reference · JavaScript

palindrome-checker

Small JavaScript testing / CI study project.

Historical testing and CI study retained as a reference; archived reference.

  • Updated: 2026-05-08
  • Language: JavaScript
Project page GitHub
Archived reference · TypeScript

real-time-analytics-stack

Event ingestion stack; archived as off-direction.

High-throughput event ingestion API with in-memory ring buffer, JSONL persistence, and WebSocket live-streaming for real-time dashboards

  • Updated: 2026-06-16
  • Language: TypeScript
Project page GitHub
Reference · archived reference · Python

super-study

AI learning and experimentation repo.

Historical AI study notes and experiments retained as a reference; archived reference.

  • Updated: 2026-05-08
  • Language: Python
Project page GitHub
Archived reference · TypeScript

WCP-Compliance-Agent-V2

Archived reference: TypeScript predecessor to WCP V5.

Superseded TypeScript predecessor. Remote archive remains after verified backup; the reference record is retained.

  • Updated: 2026-05-08
  • Language: TypeScript
Project page GitHub
Archived reference · Python

WCP-Compliance-Agent-V3

Archived reference: Python predecessor to WCP V5.

Superseded three-service predecessor. Remote archive remains after verified backup; the reference record is retained.

  • Updated: 2026-05-08
  • Language: Python
Project page GitHub
Attributed reference · archived reference

kiwi-code

Roo Code lineage retained as a reference; archived reference.

Project pageGitHub