Awesome Harness Engineering Model Card
这是可索引项目报告证据页：它保留 Awesome Harness Engineering Model Card 的源材料入口、机制线索和限制提醒；正文仍需 reader/editor 与 academic public-copy review 后才能当作最终结论引用。
Awesome Harness Engineering Model Card
| Field | Value |
|---|---|
| Repository | ai-boost/awesome-harness-engineering |
| Category | Harness Engineering 资源索引 |
| Stars / forks snapshot | 1100 / 100 |
| Language | Markdown |
| License | CC0 |
| Raw capture | raw-github/ai-boost_awesome-harness-engineering.md |
| Updated by | hourly public metadata update, 2026-06-04 01:56 +0800 |
1. Role in Self Evolve
把 harness engineering 的上下文、工具、计划、验证、记忆、沙箱、权限和观测性拆成可检索资源图。
2. Working Principle
awesome list + templates + design primitive taxonomy + adjacent collections。
3. Evidence Path
The current run rechecked a cached GitHub search result observed on 2026-06-04, with crawl recency surfaced as last week. That cached result re-confirmed the multilingual README, the harness design-primitives taxonomy, and the emphasis on memory, MCP, permissions, observability, security, evals, and templates. Exact star/fork counts were not legible in the cached snippet, so this card retains the prior 1.1k / 100 snapshot from the 2026-05-24 raw capture as the best local evidence instead of fabricating newer counts.
4. Teaching Use
Use this card to explain whether a repository is a runtime, benchmark, harness-evolution loop, memory/skill substrate, or resource index. The key reading path is: raw capture -> classification row -> public site card -> project report -> aggregate GitHub analysis.
5. Limits
这是资源索引，不是可直接运行的 agent runtime；需要和具体 harness/eval projects 联动使用。