docs: add llms.txt for agent-readable repo orientation - #48750
Closed
zuto-dev wants to merge 1 commit into
Closed
Conversation
Adds a curated llms.txt at the repo root following the llmstxt.org spec. Provides agents (Claude Code, Codex, Cursor, OpenCode, etc.) a routing index into the existing docs without re-scraping README/website/. The pitch surfaces the two architectural invariants from AGENTS.md (per-conversation prompt caching is sacred; the core is a narrow waist) so any agent reading this lands on the contribution rubric the right way around. Adds a tiny CI workflow that fails the build if any linked path in llms.txt no longer resolves, so it can't silently rot.
tonydwb
approved these changes
Jun 19, 2026
tonydwb
left a comment
There was a problem hiding this comment.
Code Review Summary
Verdict: Approved
Adds llms.txt for agent-readable repository orientation. Documentation-only addition (104 lines). No code concerns.
Reviewed by Hermes Agent
Author
|
@teknium1 friendly bump on this one — docs-only addition (104 lines, llms.txt for agent-readable repo orientation). The Hermes-Agent auto-review approved (authorAssociation NONE so it doesn't satisfy branch protection). Could one of you give it a real review when you have a sec? Happy to rebase if anything has drifted. |
Author
|
Closing — going to maintain this as a local convention for now. Thanks for taking a look, feel free to revive if useful later. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What
Adds a curated
llms.txtat the repo root following the llmstxt.org spec, plus a tiny CI workflow that fails the build if any linked path stops resolving.Why
Agents (Claude Code, Codex, Cursor, OpenCode, etc.) increasingly orient inside a codebase by reading
llms.txtfirst. Without one, they re-scrapeREADME.mdand miss thewebsite/docs/tree where the real architecture and developer-guide content lives. With one, they land on the right routing index in their first read.The pitch deliberately surfaces the two architectural invariants from
AGENTS.md— per-conversation prompt caching is sacred and the core is a narrow waist; capability lives at the edges — so any agent reading the file lands on the contribution rubric the right way around (and is less likely to propose a new core tool when a skill or plugin would do).Footprint
Per
AGENTS.md's contribution rubric:HERMES_*env vars, no plugin/core boundary changes. Pure documentation.docs-site-checks.yml.Verification
Locally:
The CI workflow runs the same check on every PR.
Notes
llms.txtper package (e.g.agent/llms.txt) is deferred until an external consumer asks.