XSkill Model Card

这是可索引项目报告证据页：它保留 XSkill 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。

XSkill Model Card

1. Role in Self Evolve

XSkill is a multimodal continual-learning paper implementation that extracts task-level skills and action-level experiences from agent trajectories, stores them in a memory bank, and re-injects them during inference across benchmark suites. It matters because self-evolving agents need explicit runtime, memory, skill, and benchmark substrates before their improvement claims become trustworthy.

2. Working Principle

collect multimodal trajectories -> summarize and critique experiences -> consolidate reusable skill documents and experience entries -> retrieve relevant memory for new tasks -> evaluate transfer on benchmark suites

3. Evidence Path

GitHub GraphQL/API snapshot captured via authenticated gh on 2026-06-13 showed 221 stars, 27 forks, 1 open issue, 0 open pull requests, 16 commits on main, latest push at 2026-05-13T03:56:12Z, no GitHub releases, and no SPDX license value returned by the API. The snapshot also showed no GitHub releases visible in the API snapshot. 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 Continual Experience and Skill Learning Paper Code: it shows how paper-code systems accumulate trajectories, distill reusable skills, and test transfer instead of only claiming continual learning in prose.

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.