object-annotation
Choose SAM3 or annotate objects yourself, then stage both for the same review.
- License 许可
- Apache-2.0
- 所属仓库
- Koooki3/LEVI
- 最近更新
- 2026-11-15
- Stars
- ★ 1
- Forks
- 0
01 前言 Front Matter
- name
- object-annotation
- metadata
- { "version": "2" }
- description
- Choose SAM3 or annotate objects yourself, then stage both for the same review.
03 SKILL.md 正文
Ask objects.strategy first: it reads this machine, not a preference. SAM3 when the worker, checkpoint and GPU headroom are there; otherwise annotate the objects yourself. Both paths stage for the same human review, so neither is a shortcut past it. Check objects.status before planning a worker run. Use the frozen run snapshot. Camera and episode scope are mandatory.
To annotate yourself, call objects.detect on the evidence you already read: it measures coherent regions and returns each one's outline, position, shape and median HSV, plus one labelled overlay. Detection measures, it does not recognise — decide from the statistics and the overlay which regions are the objects and what to call them. Submit with objects.propose citing candidate_id; send a polygon only for an object detection missed. Never invent a mask for a frame you did not read, and never name a concept outside the planned scope.
Outlines are the visible region, so a gripper covering half a plate outlines half a plate. Say so through occluded and score rather than completing the shape from memory. Report per-frame uncertainty instead of averaging it away.
04 同仓库更多 Skills
paper-to-spec
Turn a paper, preprint or technical blog into a checked method specification for LEVI: fetch the full text into an isolated directory, read the method sections verbatim, then write what the source says, what it leaves unspecified and how it maps onto the codebase. Use when the user gives a link and asks to reproduce, import or integrate a method, and also when a plan or report is about to rely on a paper's method or numbers. Not for a quick summary or a literature overview.
autoresearch
Run a measured experiment study in LEVI work: tune a harness, prompt, threshold or replay parameter against a fixed evaluator over several rounds, keep or discard each change, and confirm once on held-out data. Use when the user asks to optimise, tune, sweep, calibrate or "autoresearch" something that has a fixed evaluator and a number to move. Not for bug fixes, one-off edits or a single comparison, and never for anything that moves the real robot.
agent-brief
Write or review the prompt you hand to another model: a subagent task brief, an independent-review brief, or a prompt file for a local VLM in a LEVI harness. Use when delegating work, when writing or revising a prompt file, or when a delegated result came back wrong in a way the brief could have prevented. Not for replying to the user.
temporal-annotation
Evidence-grounded attempts, outcomes and temporal boundary review.
review-and-export
Human review and original-format export.
levi-overview
Inspect → plan → pilot → execute → validate → review → commit, at a bounded cost.