OpenSpace vs SkillX
HKUDS's skill lifecycle layer for agents: retrieve the right skill per task, evaluate which ones actually work from real outcomes, share across agents and teammates, evolve with every run. — versus — Research framework that auto-distills agent trajectories into a three-level skill knowledge base (planning, functional, atomic) — pluggable into weaker agents and new environments.
Same evolve-skills-from-experience thesis: SkillX is the research-grade trajectory distiller; OpenSpace is the product-shaped management layer with retrieval and team sharing around the same loop.
| OpenSpace | SkillX | |
|---|---|---|
| Stars | 7.1k | 265 |
| Forks | 858 | 24 |
| Language | Python | Python |
| License | MIT | MIT |
| Last activity | today | 22 days ago |
| Topics | skills | skills |
| Curated connections | 2 | 3 |
OpenSpace — the curator's take
The missing half of the skill ecosystem: everyone ships installers, almost nobody closes the loop on whether a skill WORKED. OpenSpace's bet is lifecycle — retrieval at task time, evaluation from real run outcomes, and evolution instead of repeat failures — across Claude Code, Codex, OpenClaw and friends. If your skill folder has become a junk drawer, this is the category to watch. NOT yet the standard: young v2, outcome-evaluation quality depends on how honestly your runs are scored, and the Feishu/WeChat-first community signals where the early adopters are. For plain install/audit needs, a simpler skill manager does less, more predictably.
SkillX — the curator's take
The interesting research bet: instead of storing raw trajectories or reflections, distill agent experience into a hierarchy of reusable skills that transfer — a strong backbone agent builds the library, weaker agents plug it in and improve on AppWorld/BFCL/τ2-Bench. Read it if you're building agent-improvement loops; the three-level decomposition (planning/functional/atomic) is a genuinely useful mental model. NOT production tooling — it's an academic codebase (~250 stars) built around benchmarks, so expect to adapt it to your stack rather than pip-install it.