agentleFS
NVIDIAKnown publisherNVIDIA/IsaacCapture/src/python/isaaccapture/cts/hand/AGENTS.md

IsaacCapture / hand

AGENTS.md438 starsChanged 31 days agoApache-2.0View on GitHub

CRITICAL: complete the mandatory AGENTS.md preflight in ../../../../../AGENTS.md before editing here, and read ../common/AGENTS.md and checker/AGENTS.md. - Gloves are the target. Headset optical hands cannot show 21-joint layouts, glove axis conversions or wrist-fusion faults. - Check groups are named, not numbered: signal, shape, placement, windows. - Dependency direction: capture imports the checker; the checker never imports isaaccapture or capture; the oracle imports neither the checker nor cts_common. - Physical contact is the ground truth. With no motion capture, the table,…

1 other project uses a version of this.

Adopted by

Most recently taken up by

The file

Files
AGENTS.md
<!-- SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. SPDX-License-Identifier: Apache-2.0 -->

Agent notes — cts/hand/

CRITICAL: complete the mandatory AGENTS.md preflight in ../../../../../AGENTS.md before editing here, and read ../common/AGENTS.md and checker/AGENTS.md.

  • Gloves are the target. Headset optical hands cannot show 21-joint layouts, glove axis conversions or wrist-fusion faults.
  • Check groups are named, not numbered: signal, shape, placement, windows.
  • Dependency direction: capture imports the checker; the checker never imports isaaccapture or capture; the oracle imports neither the checker nor cts_common.
  • Physical contact is the ground truth. With no motion capture, the table, pressed palms and touching fingertips are what the data can be held to; new cross-hand steps should be built on contact the performer can feel, not on a pose they approximate.
  • Thresholds come from real glove recordings only, never from optical hands or the oracle. Existing ones are loose; do not tighten one to make a fixture pass.
  • Capture pacing is a timer; nothing is pressed. A trigger moves a finger and the controller may be the wrist source. Windows are stamped with time.monotonic_ns(), which is the clock sample_time_local_common_clock is in, so labels need no record counting.
  • Name the sides in full (left, right, l_h): ruff E741 rejects a bare l, and paired left/right code reaches for it.
  • Cue WAVs are committed and hash-checked against session.py through cues/index.json; re-render with piper when wording changes.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.