# litianyi-007/harnessloop

让长时间运行的 AI agent 任务变得可审计的协议——轮次、scope-lock、证据契约、runtime 验收 eval。拒绝自相矛盾；并如实登记它强制不了什么。 · An auditability protocol for long-running AI agent tasks — rounds, scope-lock, evidence contracts, runtime acceptance evals. Refuses self-contradiction; states plainly what it cannot enforce.

## Repository

- [Catalog detail page](https://awesomeclaudeplugins.com/litianyi-007/harnessloop)
- [GitHub repository](https://github.com/litianyi-007/harnessloop)
- Stars: 0
- Forks: 0
- 1 plugin entries are reported.

## Catalog provenance

- Publication state: indexable
- Quality note: Canonical repository has a description and validated plugin count.
- Catalog snapshot: Aug 31, 2026
- Repository metadata source: public GitHub API when available

## Installation

When this repository exposes a Claude Code marketplace, add it with:

```bash
/plugin marketplace add litianyi-007/harnessloop
```