agentleFS
NVIDIAKnown publisherNVIDIA/NemoClaw/agents/langchain-deepagents-code/AGENTS.md

NemoClaw / langchain-deepagents-code

AGENTS.md23k starsChanged 13 days agoApache-2.0View on GitHub

Use the shared dependency-upgrade workflow for version changes. This guide describes the maintained component boundaries, not a release checklist. requirements.in and dependency-overrides.in feed the hash-locked dependency graph in requirements.lock. Read the lock's generation command before regenerating it. Resolve the graph on the supported Python version and target platform. Dockerfile.base installs that graph, applies dependency compatibility changes, runs pip check, and checks the runtime contract. The installed-package validator is validate-runtime-contract.py; host-side immutable-image validation lives in scripts/checks/validate-dcode-runtime-contract.mts. Dockerfile consumes the base image…

Trending

Less talk lately.about the whole repository

Dec 2025: 0 mentionsJan 2026: 0 mentionsFeb 2026: 0 mentionsMar 2026: 26 mentionsApr 2026: 6 mentionsMay 2026: 5 mentionsJun 2026: 12 mentionsJul 2026: 14 mentionsAug 2026: 1 mentionSep 2026: 2 mentionsOct 2026: 15 mentionsNov 2026: 2 mentions
Dec 2025Jun 2026Nov 2026
Mentions a month of the NVIDIA/NemoClaw repository that our weekly crawl found. There is no adoption data for this file yet.

Who's talking about NVIDIA/NemoClaw

The biggest 5 of 83 mentions linking to the repository that our crawl found. Each line is their own headline.

See all 83

Talked about on GitHub, YouTube, Bluesky, Hacker News and the web (83 times).

On GitHub · 50

45 more on GitHub

On YouTube · 21

16 more on YouTube

On Bluesky · 9

4 more on Bluesky

On Hacker News · 2

On the web · 1

The file

We split this AGENTS.md into the rules it gives an agent

10 rules. About 0 are common, 10 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈10

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use the shared dependency-upgrade workflow for version changes.

    line 6
  2. Distinctive:

    Read the lock's generation command before regenerating it.

    line 11
  3. Distinctive:

    Follow its actual COPY inputs and checksum checks when changing plugin sources or metadata.

    line 16
  4. Distinctive:

    Follow the base-image contract consumers when changing build inputs.

    line 17
  5. Distinctive:

    Remove a managed patch only when the upstream implementation and its existing behavioral regression tests establish equivalent behavior.

    line 23
  6. Distinctive:

    Keep release-specific findings in task evidence and the owning tests.

    line 24
  7. Distinctive:

    Check the host prerequisites in test guidance, including Python wheel-build packages.

    line 28
  8. Distinctive:

    Use Vitest's import graph for TypeScript consumers instead of duplicating path inventories.

    line 37
  9. Distinctive:

    Use the existing image validator on an authorized cached immutable image when installed-runtime evidence is required:

    line 42
  10. Distinctive:

    Use the current supported architecture for the artifact.

    line 49

≈ “Common” is an estimate: at least 1 in 5 of the 40,246 AGENTS.md files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
AGENTS.md
<!-- SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. --><!-- SPDX-License-Identifier: Apache-2.0 -->

Deep Agents Code Integration

Use the shared dependency-upgrade workflow for version changes. This guide describes the maintained component boundaries, not a release checklist.

Build and runtime ownership

requirements.in and dependency-overrides.in feed the hash-locked dependency graph in requirements.lock. Read the lock's generation command before regenerating it. Resolve the graph on the supported Python version and target platform. Dockerfile.base installs that graph, applies dependency compatibility changes, runs pip check, and checks the runtime contract. The installed-package validator is validate-runtime-contract.py; host-side immutable-image validation lives in scripts/checks/validate-dcode-runtime-contract.mts.

Dockerfile consumes the base image and installs managed patches and the profile plugin. Follow its actual COPY inputs and checksum checks when changing plugin sources or metadata. Follow the base-image contract consumers when changing build inputs; version searches alone miss cache and overlay dependencies.

The launchers, wrapper, and supervisor own process startup, headless output, and resume behavior. Inspect their consumers when upstream entrypoints or output formats change. Existing-sandbox version and readiness checks live in the CLI; fresh-image validation does not qualify those paths.

Remove a managed patch only when the upstream implementation and its existing behavioral regression tests establish equivalent behavior. Keep release-specific findings in task evidence and the owning tests.

Validate before image qualification

Check the host prerequisites in test guidance, including Python wheel-build packages. From the repository root, run:

npm run test:changed -- --changed=origin/main

This includes integration tests and maps opaque build inputs through test/helpers/vitest-watch-triggers.ts. Review its selected tests and unmapped inputs. Extend the existing mappings for a newly discovered opaque dependency. Use Vitest's import graph for TypeScript consumers instead of duplicating path inventories. The selected existing tests cover lock/selector agreement, patch behavior, profile-plugin checksums, image contracts, and headless consumers where those inputs changed.

These tests do not install the target wheel or qualify a real sandbox. Validate the resolved target dependency graph and changed consumers against the actual artifact before expensive qualification. Use the existing image validator on an authorized cached immutable image when installed-runtime evidence is required:

node scripts/checks/validate-dcode-runtime-contract.mts --reference '<image@sha256:digest>' --platform linux/amd64

That validator uses a read-only container without network access; Docker and the selected image must already be available. Use the current supported architecture for the artifact. Keep existing-sandbox and live lifecycle qualification separate.

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.