MCP servers and skills: search, filter by where they run, and compare.
AI-powered QA Architect agent — generates ISTQB/OWASP/IEEE/ISO-grounded Test Strategies, Risk Registers, Effort Estimates & Test Plans via RAG. Cloud-hosted (Mistral API + Pinecone).
MCPLink only★ 7
Pharos — local-first agentic RAG for your team's document library: multi-format ingest, hybrid retrieval, enterprise ACL, dual HTTP + MCP exits.
in the readmeWhat it does | |---|---|---|---| | **HTTP API** | `pharos serve` | curl, scripts, frontends | Closed-pipeline QA: `/v1/ask` runs retrieve → grounding → DeepSeek → an answer with citations, plus six retrieval endpoints
PythonMIT★ 243No key
Token-efficient e2e orchestration skill for Claude Code: explore once, replay deterministically. Playwright + axe-core + run-diff. Tests stay in your repo. MIT.
in the readmeWhat you pay for | |---|---|---|---| | **Paid SaaS** | Octomind ($89–$589/mo, now winding down ), QA Wolf (~$8k/mo entry, $60–250k/yr typical), Mabl , BrowserStack AI | Cloud parallelism, human triage
Node.jsMIT★ 4Few starsNo key
WithVibe Roadmap plugin — per-env implementation roadmap board with an MCP server the AI orchestrator drives. Apache 2.0.
in the readmeguard [~] Wire DTO validation ← active [ ] Error envelope [ ] Phase 3: Web [ ] Phase 4: QA — Plan changes — 2026-05-28 AI added Phase 4 (QA) — reason: integration tests
SkillApache-2.0Few starsNo key
Your First LLM-Wiki Conversation Knowledge Base
in the readmeLLM Wiki Auto-maintained wiki per channel — overview, topics, people, decisions, citations. QA Agent Streams cited answers over SSE. Smart router picks semantic or graph per
PythonApache-2.0★ 449No key
Shared memory MCP server — persistent, searchable, cross-client Claude, Opencode
in the readmecross-encoder reranking, NLI verification, and knowledge-graph enrichment. - **End-to-end QA:** **85.8% on the AMB harness ** (500 questions, April 2026, strict substring judge, GPT
PythonMIT★ 115No key
Local-first AI agent memory daemon. One command, works offline, no account. Markdown storage, hybrid search (SQLite FTS5 + embeddings), MCP protocol, web dashboard.
in the readmeonly * │ │ ★ Awareness Memory (Hybrid) │ 96.0% │ Hybrid RRF │ │ OMEGA │ 95.4% │ QA Accuracy │ │ Mastra (GPT-5-mini) │ 94.9% │ QA Accuracy │ │ Mastra (GPT-4o) │ 84.2% │ QA Accuracy │ │ Supermemory
Node.jsMIT★ 12No key
Zero-infrastructure, local-first memory for AI agents. One SQLite file — no Postgres, Neo4j, or cloud. Bi-temporal supersession (updates, never deletes), four-channel hybrid retrieval, and a drop-in M
in the readmecomparisons on the 0.3.x read path (September 2026: growth, footprint, matched QA) These were measured before the evidence-first engine, with real installed packages, retained
PythonApache-2.0★ 14No key
Your AI forgets everything between sessions. This fixes that — 98%+ retrieval accuracy, 100% on LongMemEval, 99% token savings. 44 MCP tools. Fully local, zero cost.
in the readmethe top-k results? This is different from end-to-end QA accuracy (retrieve + generate + judge), which is harder and lower for every system. Both measurements
Node.jsMIT★ 19No key
Who stands behind it
Who stands behind it
