Repository navigation
docs: generate llms.txt and Markdown versions of each page - #1010
jonathannorris wants to merge 2 commits into
Conversation
There was a problem hiding this comment.
Caution
Daphne (fueled by kingfisher) detected 6 potential secrets in files touched by this PR.
If any finding is a real credential (not a sample or placeholder), treat it as compromised. It's no longer confidential the moment it lands in git.
🔴 verified live · ⚪ unverified — the engine could not validate it, it may still be live
Found in touched files outside this PR's changes
Daphne scans each touched file in full, so a finding here may predate this PR — open blame to see who introduced the line.
| Secret | File |
|---|---|
| ⚪ DevCycle Server SDK Key | docs/cli-guides/environments.md:118 |
| ⚪ DevCycle Client SDK Key | docs/cli-guides/environments.md:43 |
| ⚪ DevCycle Server SDK Key | docs/cli-guides/environments.md:51 |
| ⚪ DevCycle Mobile SDK Key | docs/cli-guides/environments.md:35 |
| ⚪ DevCycle Mobile SDK Key | docs/cli-guides/environments.md:102 |
| ⚪ DevCycle Client SDK Key | docs/cli-guides/environments.md:110 |
What to do
- No need to file a Jira ticket. If a finding is confirmed live (🔴), it's already tracked confidentially in the
DTSECJira project. Tracked isn't rotated — the steps below are still yours. - Rotate or revoke it now at the provider, and update whatever consumes it.
- If this branch introduced it, purge it from history. Deleting the file or force-pushing won't remove it — it stays in every clone and the reflog. Rewrite history with
git filter-repo, then force-push. See GitHub: removing sensitive data.
Daphne alerts, it doesn't gate — it never blocks a merge or requests changes. Questions? → #dev-help-security
Deploying devcycle-docs with
|
| Latest commit: |
95c882e
|
| Status: | ✅ Deploy successful! |
| Preview URL: | https://22574a59.devcycle-docs.pages.dev |
| Branch Preview URL: | https://docs-generate-llms-txt.devcycle-docs.pages.dev |
323b8ce to
95c882e
Compare
There was a problem hiding this comment.
🔵 Needs a closer look
The custom conversion pipeline and routing workarounds need generated-output verification before final human approval.
0 open findings
What changed in this PR
Adds AI-readable exports to the Docusaurus documentation site.
Changes:
- Generates
llms.txt,llms-full.txt, and per-page Markdown with route grouping and exclusions. - Cleans converted content and repairs Markdown links.
- Corrects documentation links, formatting, and metadata.
| File | Description |
|---|---|
| yarn.lock | Locks new plugin dependencies. |
| plugins/rehype-llms-cleanup.js | Normalizes converted content and links. |
| plugins/llms-txt.js | Post-processes generated Markdown and indexes. |
| package.json | Adds the export plugin. |
| docusaurus.config.js | Configures exports and improves README heading handling. |
| docs/sdk/server-side-sdks/python/python-gettingstarted.md | Adds spacing before a wizard marker. |
| docs/sdk/server-side-sdks/node/node-typescript.md | Corrects CLI links. |
| docs/sdk/server-side-sdks/nestjs/nestjs-typescript.md | Corrects CLI links. |
| docs/sdk/client-side-sdks/react/react-typescript.md | Corrects CLI links. |
| docs/sdk/client-side-sdks/nextjs/nextjs-typescript.md | Corrects CLI links. |
| docs/sdk/client-side-sdks/javascript/javascript-typescript.md | Corrects CLI links. |
| docs/integrations/vercel-edge-config.md | Corrects SDK and quickstart links. |
| docs/integrations/index.mdx | Simplifies GitHub link wording. |
| docs/index.mdx | Marks the hero image as decorative. |
| docs/best-practices/security/compromised-keys.md | Fixes a description typo. |
| .yarnrc.yml | Supplies the plugin’s undeclared logger dependency. |
🧠 Review effort: Balanced
💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.
|
checked agaisnt https://docs-generate-llms-txt.devcycle-docs.pages.dev/llms.txt and the output looks good |
Summary
@signalwire/docusaurus-plugin-llms-txtto generatellms.txt,llms-full.txt, and a.mdcopy of each doc page, so AI tools can read the docs without scraping HTML. Same approach as feat: generate llms.txt and per-page markdown open-feature/openfeature.dev#1579./path/.mdlinks and absolute links.## Optionalinstead), the SDK example app pages, and the generateddvccommand reference, to keep the output focused.Examples
llms.txt