LightAgent Model Card
这是可索引项目报告证据页：它保留 LightAgent 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。
LightAgent Model Card
| Field | Value |
|---|---|
| Repository | wanxingai/LightAgent |
| Category | Memory/MCP Skill Agent Framework |
| Stars / forks snapshot | 1,132 / 143 |
| Commits / issues / PRs snapshot | 107 / 8 / 0 |
| Language | Python |
| License | Apache-2.0 |
| Latest visible dated signal | 2026-06-05 GitHub API snapshot |
| Raw capture | raw-github/wanxingai_lightagent.md |
| Updated by | hourly public metadata update, 2026-06-13 02:15 +0800 |
1. Role in Self Evolve
LightAgent is a lightweight Python agent framework that combines persistent memory, MCP integration, native skills, LightSwarm collaboration, and newly surfaced LightFlow workflow orchestration into a small-footprint self-learning runtime. It matters because self-evolving agents need explicit runtime, memory, skill, and benchmark substrates before their improvement claims become trustworthy.
2. Working Principle
compose lightweight agents with tools, MCP, and memory -> add native skills and optional trace observability -> delegate via LightSwarm -> chain deterministic multi-step flows with LightFlow -> keep self-learning behavior grounded in runtime memory and reusable tool plans
3. Evidence Path
GitHub GraphQL/API snapshot captured via authenticated gh on 2026-06-13 showed 1,132 stars, 143 forks, 8 open issues, 0 open pull requests, 107 commits on main, latest push at 2026-06-05T15:00:09Z, Apache-2.0 license, and latest release LightAgent v0.8.0 published on 2026-06-05. The snapshot also showed 21 releases with LightAgent v0.8.0 latest on 2026-06-05. This run keeps freshness honest because it uses authenticated GitHub API data rather than stale local summaries.
4. Teaching Use
Use this card to explain Memory/MCP Skill Agent Framework: it shows how lightweight runtimes attach memory, MCP, skills, and workflow layers without becoming a monolithic product stack.
5. Limits
The repository was not cloned in this iteration; no benchmark run, workflow execution, or agent loop experiment was executed. Counts and claims are GitHub API snapshot signals unless independently revalidated later.