Skip to content

[None][docs] Fix AGENTS.md accuracy and reduce context bloat#12258

Merged
kaiyux merged 2 commits into
NVIDIA:mainfrom
kaiyux:fix-agents-md-accuracy
Mar 18, 2026
Merged

[None][docs] Fix AGENTS.md accuracy and reduce context bloat#12258
kaiyux merged 2 commits into
NVIDIA:mainfrom
kaiyux:fix-agents-md-accuracy

Conversation

@kaiyux
Copy link
Copy Markdown
Member

@kaiyux kaiyux commented Mar 16, 2026 •
edited
Loading

Summary

  • Fix stale/incorrect claims in AGENTS.md: wrong api_stability path, deprecated backend parameter shown as primary API, non-existent ADExecutor class, wrong parent class name (LlmArgs → BaseLlmArgs), outdated config count
  • Add missing info: _torch/models/ directory, attention backends, CODING_GUIDELINES.md in docs table
  • Extract 73-line CI failure retrieval scripts to .github/ci-failure-retrieval.md to reduce AGENTS.md size by 28%

Test plan

  • Verify all referenced file paths exist in the repo
  • Confirm class/API names match current codebase
  • Check that .github/ci-failure-retrieval.md content matches what was removed from AGENTS.md

🤖 Generated with Claude Code

Summary by CodeRabbit

  • Documentation
    • Added guide for retrieving and diagnosing CI test failures from pull requests
    • Updated backend configuration documentation with new argument-based signatures and deprecation notes for legacy approaches
    • Expanded documentation with additional topics including AutoDeploy, quantization strategies, and deployment guides

@kaiyux kaiyux requested review from a team as code owners March 16, 2026 23:26
kaiyux
kaiyux commented Mar 16, 2026
View reviewed changes
Comment thread AGENTS.md
@coderabbitai
Copy link
Copy Markdown
Contributor

coderabbitai Bot commented Mar 16, 2026 •
edited
Loading

📝 Walkthrough

Walkthrough

This PR adds a new CI failure retrieval guide and updates the AGENTS.md documentation to reflect revised backend configurations, updated design pattern hierarchies, expanded key documentation references, and refined entry point signatures for PyTorch and TensorRT backends.

Changes

Cohort / File(s) Summary
CI Documentation
.github/ci-failure-retrieval.md		 New documentation explaining multi-step process for retrieving and diagnosing CI test failures from PRs using GitHub API and Jenkins testReport API, with details on failure fields and optional stdout/stderr retrieval.
Agents Documentation
AGENTS.md		 Updated backend configurations with new argument-based signatures (TorchLlmArgs, TrtLlmArgs), revised design pattern hierarchy (BaseLlmArgs as base class), clarified attention backends, expanded key documentation topics, updated API stability test references, and refined CI failure retrieval guidance.

Estimated code review effort

🎯 3 (Moderate) | ⏱️ ~20 minutes

🚥 Pre-merge checks | ✅ 3
✅ Passed checks (3 passed)
Check name Status Explanation
Title check ✅ Passed The title accurately describes the main changes: fixing accuracy issues in AGENTS.md and reducing its size by extracting CI documentation.
Docstring Coverage ✅ Passed No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Description check ✅ Passed The PR description clearly explains the changes, test plan, and objectives. It covers what was fixed, what was added, and how it was validated.

✏️ Tip: You can configure your own custom pre-merge checks in the settings.

✨ Finishing Touches
🧪 Generate unit tests (beta)
  • Create PR with unit tests
📝 Coding Plan
  • Generate coding plan for human review comments

Comment @coderabbitai help to get the list of available commands and usage tips.

Tip

CodeRabbit can scan for known vulnerabilities in your dependencies using OSV Scanner.

OSV Scanner will automatically detect and report security vulnerabilities in your project's dependencies. No additional configuration is required.

venkywonka
venkywonka approved these changes Mar 17, 2026
View reviewed changes
Copy link
Copy Markdown
Collaborator

@venkywonka venkywonka left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

thank you @kaiyux - although this is out of scope of this pr - if you could just make ci-failure-retreival.md into a skill and put it in .claude/skills then that would be the cherry-on-top :D
feel free to ignore - this is great as-is!

venkywonka
venkywonka reviewed Mar 17, 2026
View reviewed changes
Comment thread AGENTS.md Outdated
kaiyux added 2 commits March 17, 2026 16:18
@kaiyux
[None][docs] Fix AGENTS.md accuracy and reduce context bloat
3e5cf75 
Fix stale/incorrect claims: wrong api_stability path, deprecated backend
parameter shown as primary API, non-existent ADExecutor class, wrong parent
class name (LlmArgs → BaseLlmArgs), outdated config count. Add missing info:
_torch/models/ directory, attention backends, CODING_GUIDELINES.md in docs
table. Extract 73-line CI failure retrieval scripts to .github/ci-failure-retrieval.md
to reduce AGENTS.md size by 28%.

Signed-off-by: Kaiyu Xie <[email protected]>
@kaiyux
[None][docs] Address PR review: remove config count, add CI retrieval…
2d186b8 
… skill

Remove hardcoded config count from AGENTS.md to avoid maintenance overhead.
Move ci-failure-retrieval from .github/ to .claude/skills/ as a proper
Claude Code skill with YAML frontmatter, matching the format of other
skills like ad-pipeline-failure-pr.

Signed-off-by: Kaiyu Xie <[email protected]>
@kaiyux kaiyux force-pushed the fix-agents-md-accuracy branch from 4a9d01f to 2d186b8 Compare March 17, 2026 23:18
@kaiyux kaiyux enabled auto-merge (squash) March 18, 2026 05:31
@kaiyux
Copy link
Copy Markdown
Member Author

kaiyux commented Mar 18, 2026

/bot skip --comment "doc changes"

@tensorrt-cicd
Copy link
Copy Markdown
Collaborator

tensorrt-cicd commented Mar 18, 2026

PR_Github #39393 [ skip ] triggered by Bot. Commit: 2d186b8 Link to invocation

@tensorrt-cicd
Copy link
Copy Markdown
Collaborator

tensorrt-cicd commented Mar 18, 2026

PR_Github #39393 [ skip ] completed with state SUCCESS. Commit: 2d186b8
Skipping testing for commit 2d186b8

Link to invocation

@kaiyux kaiyux merged commit 8a723f9 into NVIDIA:main Mar 18, 2026
10 of 11 checks passed
limin2021 pushed a commit to limin2021/TensorRT-LLM that referenced this pull request Mar 19, 2026
@kaiyux
[None][docs] Fix AGENTS.md accuracy and reduce context bloat (NVIDIA#…
b387cbb 
…12258)

Signed-off-by: Kaiyu Xie <[email protected]>
longcheng-nv pushed a commit to longcheng-nv/TensorRT-LLM that referenced this pull request Mar 31, 2026
@kaiyux @longcheng-nv
[None][docs] Fix AGENTS.md accuracy and reduce context bloat (NVIDIA#…
6e345c5 
…12258)

Signed-off-by: Kaiyu Xie <[email protected]>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@venkywonka venkywonka venkywonka approved these changes

@arysef arysef Awaiting requested review from arysef arysef is a code owner automatically assigned from NVIDIA/trt-llm-doc-owners

@laikhtewari laikhtewari Awaiting requested review from laikhtewari laikhtewari is a code owner automatically assigned from NVIDIA/trt-llm-doc-owners

@niukuo niukuo Awaiting requested review from niukuo niukuo was automatically assigned from NVIDIA/trt-llm-gh-workflows-infra-devs

@zeroepoch zeroepoch Awaiting requested review from zeroepoch zeroepoch was automatically assigned from NVIDIA/trt-llm-infra-devs

Assignees

@kaiyux kaiyux

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

3 participants

@kaiyux @tensorrt-cicd @venkywonka