skills.rest
Arize-ai
Arize-aiOFFICIAL
Explore all skills by this author →
Website𝕏@arizeaiGitHub

phoenix-llms-txt

Maintains and audits the Phoenix llms.txt machine-readable documentation index.

11.3k|1.1k|Updated Nov 9, 2022
Search Tags:#llms.txt#Documentation#Phoenix#Mintlify#docs index#coverage audit#AI Agents
One-click install
npx skills add https://github.com/Arize-ai/phoenix --skill phoenix-llms-txt
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: phoenix-llms-txt
Source: https://github.com/Arize-ai/phoenix/tree/main/.agents/skills/phoenix-llms-txt
Command: npx skills add https://github.com/Arize-ai/phoenix --skill phoenix-llms-txt

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Keeping the Phoenix llms.txt documentation index accurate and complete is tedious: pages get added, renamed, or removed, and coverage gaps cause AI agents and the px docs fetch CLI to miss important docs. This Skill provides the rules and audit workflow to keep the index synchronized with the published docs tree.

Core Features & Use Cases

  • Coverage Auditing: Runs a five-step diff workflow that intersects docs.json navigation with on-disk .mdx files, then compares against current llms.txt entries to find missing and stale URLs.
  • Editorial Rules: Defines exactly which pages to include (workflows, SDK references, integrations, release notes) and exclude (notebooks, demos, translated pages, orphan files), plus section ordering and title/description writing standards.
  • Verification: Validates changes by running the phoenix-cli docs parser tests to confirm the file remains well-formed.
  • Use Case: After adding a new evaluation metrics page to the Phoenix docs, use this Skill to add a properly titled and described llms.txt entry, audit overall coverage against the 90% target, and verify the file parses correctly.

Quick Start

Audit the Phoenix llms.txt file for missing or stale documentation entries and update it to full coverage.

Frequently Asked Questions about phoenix-llms-txt

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I audit llms.txt coverage against published docs?▼

Intersect the page paths in docs.json navigation with .mdx files on disk to get published URLs, extract URLs currently in llms.txt, then diff the two sets with comm to find missing and stale entries. Target at least 90% coverage of published pages.

What pages should be excluded from an llms.txt index?▼

Exclude interactive demos, Colab notebook links, cookbook pages that are only embeds, translated pages, draft or orphan .mdx files not in navigation, bare external links, and duplicate entries. These provide no usable content for LLM consumers.

How do I write good llms.txt entry descriptions?▼

Write action-oriented descriptions of 5-20 words using imperative verbs like Configure, Instrument, or Run. The description must add information beyond the title, mention package names for SDK pages, and avoid filler words like comprehensive or learn more.

How do I verify llms.txt changes are well-formed?▼

Run the phoenix-cli docs parser tests with pnpm test -- --grep "docs" from the js/packages/phoenix-cli directory. This confirms the file parses correctly for the px docs fetch CLI.

Why do some llms.txt URLs appear stale but should be kept?▼

Non-filesystem resources like the OpenAPI spec URL and external TypeScript SDK docs have no corresponding .mdx file, so they show up in the stale diff. These are expected results and should be kept because they remain valid machine-readable resources.

Related Skills

Explore More Software Engineering Skills→
ophidiariumophidiarium
5

mintlify

Build Mintlify documentation sites by authoring MDX pages and configuring docs.json navigation.

Official
Intermediate
databricks-solutionsdatabricks-solutions
1.8k

databricks-docs

Retrieve Databricks documentation pages and the llms.txt index.

Official
Basic
Hung-3008Hung-3008
1

scientific-schematics

Generate publication-quality scientific diagrams from natural-language prompts with iterative refinement.

Community
Advanced
joelhooksjoelhooks
22

vectorize-search

Index plaintext content and store embeddings for semantic search over encrypted memories.

Community
Advanced