AI Agent Trust Snapshot — Week 25, 2026
Every week HVTracker diffs its registry snapshots and publishes what actually changed: trust-score moves of 3+ points, new listings, and gained provenance or MCP support. This edition compares the 2026-06-14 and 2026-06-21 snapshots — 24 gains, 18 declines, 78 new listings, 0 provenance and 1 MCP additions. Every figure below is generated directly from those snapshots.
Trust score up (24)
| Agent | HVTrust | Change |
|---|---|---|
| ASM | 22.8 → 59.4 | +36.6 |
| TEN Framework | 24.0 → 60.3 | +36.3 |
| Obsidian Second Brain | 23.9 → 60.0 | +36.1 |
| Engram | 24.9 → 60.4 | +35.5 |
| Hivemind | 23.9 → 57.1 | +33.2 |
| TencentDB Agent Memory | 24.0 → 56.1 | +32.1 |
| Turbovec | 24.4 → 56.1 | +31.7 |
| PraisonAI | 38.1 → 67.2 | +29.1 |
| AI Town | 21.4 → 49.9 | +28.5 |
| SWE-agent | 37.9 → 66.1 | +28.2 |
| HyperAgent | 19.6 → 46.7 | +27.1 |
| Burr | 38.6 → 65.1 | +26.5 |
| Griptape | 38.0 → 64.4 | +26.4 |
| ClawWork | 17.6 → 41.8 | +24.2 |
| OpenHands | 51.2 → 71.6 | +20.4 |
| Roo Code | 35.4 → 54.7 | +19.3 |
| Marvin | 36.3 → 54.8 | +18.5 |
| Aider | 39.0 → 56.3 | +17.3 |
| MetaGPT | 30.2 → 44.6 | +14.4 |
| Superpowers | 59.6 → 67.7 | +8.1 |
| Guardrails AI | 80.2 → 86.8 | +6.6 |
| Semantic Kernel | 67.1 → 71.6 | +4.5 |
| pgvector | 52.8 → 56.9 | +4.1 |
| Open Interpreter | 68.7 → 72.4 | +3.7 |
Trust score down (18)
| Agent | HVTrust | Change |
|---|---|---|
| CopilotKit | 76.7 → 45.7 | -31.0 |
| Bernstein | 81.4 → 75.8 | -5.6 |
| BrowserGym agents | 57.7 → 52.2 | -5.5 |
| PydanticAI | 90.2 → 85.1 | -5.1 |
| Skyvern | 88.5 → 83.4 | -5.1 |
| agentmemory | 71.0 → 65.9 | -5.1 |
| Promptfoo | 76.8 → 71.8 | -5.0 |
| Qdrant | 74.5 → 69.5 | -5.0 |
| OpenClaw | 88.7 → 84.2 | -4.5 |
| OpenAI Agents SDK | 88.9 → 84.5 | -4.4 |
| Swarm by OpenAI | 52.1 → 47.9 | -4.2 |
| Quivr | 45.9 → 41.8 | -4.1 |
| AutoGPT | 69.6 → 65.9 | -3.7 |
| LlamaIndex | 46.0 → 42.3 | -3.7 |
| Rasa | 60.7 → 57.1 | -3.6 |
| Sotopia agents | 51.3 → 47.8 | -3.5 |
| Agent-S | 49.9 → 46.7 | -3.2 |
| Qwen Agent | 51.2 → 48.1 | -3.1 |
Newly listed (78)
| Agent | Category | HVTrust |
|---|---|---|
| Hindsight | Memory & Knowledge | 85.2 |
| Threadplane | Agent Frameworks | 82.2 |
| MCP Python SDK | Protocols & Tool Integration | 70.1 |
| Zvec | Memory & Knowledge | 68.7 |
| AG2 | Agent Frameworks | 68.1 |
| Context7 | Protocols & Tool Integration | 66.8 |
| Microsoft MCP Catalog | Protocols & Tool Integration | 66.8 |
| MCP TypeScript SDK | Protocols & Tool Integration | 66.2 |
| mcp-use | Protocols & Tool Integration | 63.6 |
| Ponytail | Coding Agents | 62.5 |
| Eve | Needs review | 61.5 |
| Serena | Protocols & Tool Integration | 60.8 |
| Jido | Agent Frameworks | 60.3 |
| Agent Reach | Protocols & Tool Integration | 58.8 |
| Lumibot | Research & Data | 56.3 |
| Habitat-Lab | Robotics & Embodied | 52.6 |
| autoMate | Browser & Computer Use | 48.1 |
| YoMo | Agent Frameworks | 45.7 |
| Srcbook | UI & App Builders | 45.6 |
| Chat Nio | LLM Gateways & Infra | 43.3 |
| Open Deep Research (dzhng) | Research & Data | 42.7 |
| Youtu-Agent | Agent Frameworks | 41.3 |
| MCP Playwright | Protocols & Tool Integration | 36.2 |
| ScaleCUA | Browser & Computer Use | 33.6 |
| PyWinAssistant | Browser & Computer Use | 29.5 |
| Dyad | UI & App Builders | 27.2 |
| AstrBot | Voice & Conversational | 26.9 |
| Chrome DevTools MCP | Protocols & Tool Integration | 26.9 |
| MCP Specification | Protocols & Tool Integration | 26.8 |
| Hatchet | Workflow Platforms | 26.8 |
| ToolJet | UI & App Builders | 26.7 |
| Omnigent | Agent Frameworks | 26.5 |
| Open Multi-Agent | Multi-Agent Systems | 26.2 |
| GitHub MCP Server | Protocols & Tool Integration | 26.0 |
| Conductor | Workflow Platforms | 26.0 |
| DeepSeek-Reasonix | Coding Agents | 25.6 |
| n8n-MCP | Protocols & Tool Integration | 25.6 |
| Figma Context MCP | Protocols & Tool Integration | 25.4 |
| LazyLLM | Agent Frameworks | 25.4 |
| OpenLIT | Observability & Evaluation | 25.4 |
| Golem | Workflow Platforms | 24.6 |
| fast-agent | Agent Frameworks | 24.4 |
| OpenBB | Research & Data | 24.2 |
| Steel Browser | Browser & Computer Use | 24.0 |
| IntentKit | Agent Frameworks | 23.8 |
| MCP Go | Protocols & Tool Integration | 23.6 |
| Agently | Agent Frameworks | 23.2 |
| Stakpak Agent | Coding Agents | 22.9 |
| Agentic Security | Security & Guardrails | 22.7 |
| BotSharp | Agent Frameworks | 22.7 |
| Yao | UI & App Builders | 22.4 |
| Agent Squad | Multi-Agent Systems | 22.3 |
| HumanLayer | Coding Agents | 22.1 |
| LLM Wiki Agent | Memory & Knowledge | 22.0 |
| Upsonic | Agent Frameworks | 21.9 |
| PentestAgent | Security & Guardrails | 21.8 |
| LiteFlow | Workflow Platforms | 21.7 |
| Mirascope | Agent Frameworks | 21.7 |
| Kodu | Coding Agents | 21.7 |
| ByteRover CLI | Memory & Knowledge | 21.5 |
| FinRobot | Research & Data | 21.0 |
| Parlant | Agent Frameworks | 20.9 |
| TuriX | Browser & Computer Use | 20.6 |
| Call Center AI | Voice & Conversational | 20.6 |
| Convex Chef | UI & App Builders | 20.5 |
| MiroThinker | Research & Data | 19.7 |
| agentUniverse | Agent Frameworks | 19.1 |
| HexStrike AI | Security & Guardrails | 18.7 |
| PocketFlow | Agent Frameworks | 18.6 |
| Tongyi DeepResearch | Research & Data | 18.5 |
| DeepResearchAgent | Multi-Agent Systems | 18.4 |
| Magentic | Agent Frameworks | 18.2 |
| OpenAI CUA Sample App | Browser & Computer Use | 18.0 |
| OpenBrowser | Browser & Computer Use | 17.8 |
| AgentBench | Observability & Evaluation | 16.7 |
| ChatArena | Multi-Agent Systems | 16.0 |
| Nanobrowser | Browser & Computer Use | 15.3 |
| AutoAgent | Agent Frameworks | 15.3 |
MCP support gained (1)
| Agent | Category | Status |
|---|---|---|
| Vellum Assistant | Memory & Knowledge | declared |
Generated automatically from the 2026-06-14 and 2026-06-21
registry snapshots — no numbers in this post are estimated or written by hand.
Method: HVTrust methodology ·
live diff: what changed this week ·
feed: RSS.