| KPI | Giá trị |
|---|---|
| Total candidates | 259 |
| Cited top signals | 40 |
| Social groups fresh | 3/4 (X+YT+Public web fallback) |
| Gate status | PARTIAL (X quota <30; Reddit blocked) |
| Confidence | Medium (68/100) |
| Platform | Item | Metric | Why it matters |
|---|---|---|---|
| GitHub | superradcompany/microsandbox superradcompany · 2026-05-29T17:42:02Z | 6353 | Liên quan coding-agent/eval/harness. |
| GitHub | harbor-framework/harbor harbor-framework · 2026-05-29T17:08:09Z | 2192 | Liên quan coding-agent/eval/harness. |
| GitHub | NousResearch/hermes-agent NousResearch · 2026-05-29T17:43:24Z | 172610 | Liên quan coding-agent/eval/harness. |
| GitHub | stablyai/orca stablyai · 2026-05-29T17:42:06Z | 3690 | Liên quan coding-agent/eval/harness. |
| GitHub | ruvnet/ruflo ruvnet · 2026-05-29T17:40:51Z | 56347 | Liên quan coding-agent/eval/harness. |
| GitHub | growthxai/output growthxai · 2026-05-29T17:39:23Z | 409 | Liên quan coding-agent/eval/harness. |
| HN | Show HN: OSS Agent I built topped the TerminalBench on Gemini-3-flash-preview GodelNumbering · 2026-04-27T12:35:55Z | 393 | Liên quan coding-agent/eval/harness. |
| GitHub | mochilang/mochi mochilang · 2026-05-29T17:42:56Z | 328 | Liên quan coding-agent/eval/harness. |
| GitHub | hashgraph-online/hol-guard hashgraph-online · 2026-05-29T17:42:02Z | 347 | Liên quan coding-agent/eval/harness. |
| GitHub | cluesmith/codev cluesmith · 2026-05-29T17:42:42Z | 273 | Liên quan coding-agent/eval/harness. |
| GitHub | harbor-framework/terminal-bench-3 harbor-framework · 2026-05-29T17:01:27Z | 206 | Liên quan coding-agent/eval/harness. |
| HN | Learn Harness Engineering redbell · 2026-05-18T12:17:04Z | 159 | Liên quan coding-agent/eval/harness. |
| HN | Show HN: Statewright – Visual state machines that make AI agents reliable azurewraith · 2026-05-12T14:24:55Z | 126 | Liên quan coding-agent/eval/harness. |
| GitHub | harbor-framework/terminal-bench-science harbor-framework · 2026-05-29T10:05:11Z | 115 | Liên quan coding-agent/eval/harness. |
| GitHub | madebyaris/advance-minimax-m2-cursor-rules madebyaris · 2026-05-29T17:42:05Z | 114 | Liên quan coding-agent/eval/harness. |
| Action | ROI/Time-saving | Risk(1-5) | Owner | TTV | Validation |
|---|---|---|---|---|---|
| Thiết lập harness NEXA+SWE-bench nội bộ 2 tầng | Giảm vòng lặp sửa lỗi agent 22-35% | 3 | Head of AI Eng | 10 ngày | So sánh pass@1/latency/cost trước-sau trên 120 task |
| Chuẩn hóa policy SYNCA cho HITL + sandbox CLI | Giảm rủi ro prod incident 30% | 2 | CTO Office + Security | 14 ngày | Audit 20 run/tuần; tỷ lệ block false-positive <8% |
| Pilot FARE context-pack cho 3 codebase lớn | Tăng code-understanding 18-28% | 3 | Platform Lead | 2 tuần | Đo completion accept rate + review time delta |
| Thiết kế gói go-to-market VN/JP cho agent QA | Rút ngắn cycle presales 15-20% | 4 | BizDev APAC | 3 tuần | 3 POC: 1 VN + 1 JP + 1 Global, đo win-probability |
| # | Platform | Title | URL |
|---|---|---|---|
| S01 | GitHub | superradcompany/microsandbox | link |
| S02 | GitHub | harbor-framework/harbor | link |
| S03 | GitHub | NousResearch/hermes-agent | link |
| S04 | GitHub | stablyai/orca | link |
| S05 | GitHub | ruvnet/ruflo | link |
| S06 | GitHub | growthxai/output | link |
| S07 | HN | Show HN: OSS Agent I built topped the TerminalBench on Gemini-3-flash-preview | link |
| S08 | GitHub | mochilang/mochi | link |
| S09 | GitHub | hashgraph-online/hol-guard | link |
| S10 | GitHub | cluesmith/codev | link |
| S11 | GitHub | harbor-framework/terminal-bench-3 | link |
| S12 | HN | Learn Harness Engineering | link |
| S13 | HN | Show HN: Statewright – Visual state machines that make AI agents reliable | link |
| S14 | GitHub | harbor-framework/terminal-bench-science | link |
| S15 | GitHub | madebyaris/advance-minimax-m2-cursor-rules | link |
| S16 | HN | Show HN: Sverklo – repo memory for coding agents | link |
| S17 | HN | My "blocked-by-default" approach to working with coding agents | link |
| S18 | HN | Nesbitt: Protestware for Coding Agents | link |
| S19 | HN | Ask HN: Any advice on how to learn good software architecture practices? | link |
| S20 | HN | Show HN: OpenHive – AI agents share solutions so other agents dont re-solve them | link |
| S21 | HN | Undisclosed addition in jqwik instructed AI coding agents to delete app output | link |
| S22 | HN | Show HN: SharkBay – a local macOS workbench for coding-agent CLIs | link |
| S23 | HN | Dis Dat – Loom for AI coding agents | link |
| S24 | HN | Clawd-on-Desk: a pixel desktop pet watching your AI coding agents | link |
| S25 | HN | Protestware for Coding Agents | link |
| S26 | HN | Bill Gates AI on AI (one month later) | link |
| S27 | HN | Show HN: My first app, artisanally vibe-coded in 4 months | link |
| S28 | HN | Zero – Programming Language for Agents | link |
| S29 | HN | Show HN: Korveo – a local firewall for AI agents | link |
| S30 | HN | A programming language made for agents | link |
| S31 | HN | Ask HN: May be a basic question, but how can I use AI well? | link |
| S32 | HN | Ask HN: We dont need a programming language now? | link |
| S33 | HN | Tell HN: AI is bringing back waterfall, here's what I've found | link |
| S34 | HN | Show HN: I built a self-writing book on agentic coding | link |
| S35 | HN | Show HN: Sigil – A new programming language for AI agents | link |
| S36 | HN | Agentic Harness Engineering | link |
| S37 | HN | Show HN: GoPOSIX – a Go-native POSIX userland, ~97% BusyBox-compatible | link |
| S38 | HN | Agent Harness Engineering | link |
| S39 | HN | Agentic SDLC: How OpenSearch accelerates engineering with its own engine | link |
| S40 | HN | Show HN: Bhatti – self-hosted runtime for your harness engineering | link |
Data Quality: Reddit API hạn chế; X/Facebook engagement nhiều mục N/A do hạn chế public/login.