Skip to content

docs: Add AGENTS.md and llms.txt#2797

Draft
Ly-Joey wants to merge 5 commits into
google:mainfrom
Ly-Joey:docs-agents-md
Draft

docs: Add AGENTS.md and llms.txt#2797
Ly-Joey wants to merge 5 commits into
google:mainfrom
Ly-Joey:docs-agents-md

Conversation

@Ly-Joey
Copy link
Copy Markdown
Contributor

@Ly-Joey Ly-Joey commented May 12, 2026

(Still WIP but comments are very welcomed!)

Attempt at:

  • llms.txt: outline project structure and acts as a sitemap for LLMs
  • AGENTS.md: added instruction / context for agents

I'm inclined to test this out a bit more over the week before merging into main branch.

@Ly-Joey Ly-Joey requested review from another-rex and cuixq May 13, 2026 04:35
Copy link
Copy Markdown
Collaborator

@another-rex another-rex left a comment

Choose a reason for hiding this comment

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

Hmm I was thinking this would also include how osv-scanner and osv-scalibr works together and the plugin system...etc

Not sure if that fits into this file though

Comment thread AGENTS.md

Before requesting a human review on a Pull Request, ensure that the following checks pass:

- **Linting:** Run `./scripts/run_lints.sh` and resolve all warnings and errors.
Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

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

Suggested change
- **Linting:** Run `./scripts/run_lints.sh` and resolve all warnings and errors.
- **Linting:** Run `./scripts/run_lints.sh` and resolve all warnings and errors. If you run into a toolchain error about go being tool old. Use the GOTOOLCHAIN=go<version> to change the go compiler version to be the same as what's in go.mod. Example `GOTOOLCHAIN=go1.26.3 ./scripts/run_lints.sh`

Comment thread AGENTS.md Outdated
Comment on lines +20 to +22
- **Strict Issue-First Workflow:** You must follow the issue-first workflow before opening any Pull Request. The required path is always: **Issue Assignment -> Pull Request**.
- **Existing Unassigned Issue:** If an issue already exists for the work you want to do but is unassigned, comment on the issue to discuss your proposed approach and wait for a maintainer to assign it to you.
- **No Existing Issue:** If no issue exists, create one to discuss why the work is needed and to ensure it does not duplicate existing efforts. Wait for the issue to be assigned to you before proceeding.
Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

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

technically this does not cover "an existing issue that is assigned to someone else" - I suspect LLMs might be able to infer that enough that it means we won't get tons of PRs via that as a loophole, but still something we might as well cover 🤷

@cuixq
Copy link
Copy Markdown
Contributor

cuixq commented May 13, 2026

Hmm I was thinking this would also include how osv-scanner and osv-scalibr works together and the plugin system...etc

Not sure if that fits into this file though

Yeah I also think it will be helpful if we could explain in more details how osv-scanner works with osv-scalibr, so that if there is an issue with scalibr, the agent knows to open an issue or make contributions there.

Comment thread AGENTS.md Outdated

### 3.3 Commit Messages and PR Titles

- **Conventional Commits:** You must follow the [Conventional Commits](https://www.conventionalcommits.org/) specification for all pull request titles and commit messages.
Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

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

also mention that we have a section for that in https://github.com/google/osv-scanner/blob/main/CONTRIBUTING.md#making-commits as well

Comment thread AGENTS.md Outdated

- **Coverage:** Any new feature, bug fix, or refactor should be accompanied by relevant tests (unit, integration, or snapshot tests).
- **Snapshots:** If your changes modify behavior that affects snapshot tests, update the snapshots (e.g., using `make test SNAPS=true`) and explicitly note this in the PR description.
- **VCR Cassettes:** If your changes add new HTTP interactions in tests that use `go-vcr`, ensure the appropriate cassettes are recorded or updated as described in `CONTRIBUTING.md`.
Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

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

maybe also mention that Makefile has the definition as well, and it can also be useful to understand the use cases for other testing usage.

Comment thread AGENTS.md Outdated
Co-authored-by: Gareth Jones <3151613+G-Rath@users.noreply.github.com>
@codecov-commenter
Copy link
Copy Markdown

codecov-commenter commented May 15, 2026

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 79.14%. Comparing base (408fcd6) to head (9e88362).
⚠️ Report is 6 commits behind head on main.

Additional details and impacted files
@@            Coverage Diff             @@
##             main    #2797      +/-   ##
==========================================
- Coverage   79.19%   79.14%   -0.05%     
==========================================
  Files         121      121              
  Lines        8185     8185              
==========================================
- Hits         6482     6478       -4     
- Misses       1322     1324       +2     
- Partials      381      383       +2     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:
  • ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
  • 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

5 participants