Localize objects and compose low-level CAP MCP actions using observations, Cartesian or synchronized motion, joint targets, and single-arm gripper control.
中文简介正在准备中。当前显示原文说明。
NVlabs/ASPIRE8872026年10月16日 更新
免费在 GitHub 公开
manipulation-tasks
Perform task-specific real YAM manipulation with native CAP tools. Use for solving or turning a physical Rubik's Cube, bottle-cap removal and pouring, pill insertion and bottle closure, or inverted cup nesting.
中文简介正在准备中。当前显示原文说明。
NVlabs/ASPIRE星标 887(GitHub 887 + cccskills 0)
实测评分 未实测
查看安装方法包含的文件（2）
- SKILL.md1.4 KB
- rubiks-cube.md15.0 KB
SKILL.md（原文）
安装前可以查看交给 AI 代理的指令内容。
Manipulation tasks
Follow AGENTS.md and the loaded CAP operational skills. Read the guide for the requested task:
- Rubik's Cube: reconstruct the cube, execute supported layer turns, reorient, and place the solved cube. Includes the reusable solve cycle.
Read the cube task context for the requested outcome and the recipe for a new scene.
Other task recipes (read only the one matching the operator's task):
- Unscrew and pour: stabilize the table-supported bottle, turn and regrasp the cap, verify detachment, and pour onto the table.
- Pill insertion and bottle closure: acquire the pill, verify entry, seat the cap, and distinguish threaded seating from unverified final tightness.
- Inverted cup nesting: count all cups, orient the opening downward, establish insertion before release, and recover without disturbing the existing stack.
评价
还没有评价。欢迎分享你的使用感受。
评论
无需登录即可发表。欢迎分享使用感受、提问，或说明在哪些环境下无法运行。
还没有评论。来写第一条吧。
同一仓库的技能
Run the CaP Fix Loop + Evolutionary Search experiment on a user-supplied list of low-performing benchmark tasks: iterative K=8 candidate search on development seeds 51–65, validation selection, and final held-out evaluation on seeds 1–50.
中文简介正在准备中。当前显示原文说明。
NVlabs/ASPIRE8872026年10月16日 更新
Run the baseline-free CaP iterative debugging (fix loop) pipeline on a user-supplied list of benchmark tasks: inspect one initial observed scene per task, generate task-level code, debug failures on development seeds using recorded run artifacts, validate on the held-out partition, grade the skill library against what actually worked, and promote only those patterns a verifier could reproduce the fix from.
中文简介正在准备中。当前显示原文说明。
NVlabs/ASPIRE8872026年10月16日 更新
Block-by-block fix loop on BEHAVIOR-1K (R1 Pro, OmniGibson on Isaac Sim). The policy is grown one code block at a time against what the robot actually did, and what carries across seeds is a skill library, not a program. Use for the two R1 Pro pickup tasks; Stage 1 learns on seeds 26-35, Stage 2 evaluates on 1-25 with a frozen library.
中文简介正在准备中。当前显示原文说明。
NVlabs/ASPIRE8872026年10月16日 更新