arize-ai/phoenix
23 skills3.7k installs248k starsGitHub
Install
npx skills add https://github.com/arize-ai/phoenixSkills in this repo
1Phoenix Cliphoenix-cli is an Arize Phoenix agent skill for developers debugging instrumented LLM and agent applications who have qualitative failure notes but need quantitative structure. Its axial-coding reference groups open-ended observations—trace notes, span reviews, or open-coding output—into named MECE categories with counts grounded in real traffic rather than invented top-down labels. The workflow pairs with open coding via shared session identifiers and Phoenix CLI commands such as px trace add-note and px span annotate, plus GraphQL API access for session rollups. Reach for phoenix-cli when asking what failure categories exist, which evals to build next, or how to prioritize fixes after traces are already collected. The skill supports trace-, span-, and session-level units of analysis so multi-turn agent trajectory failures roll up correctly before eval design. Recommended flow runs open-coding notes first, then axial grouping, then eval construction for the highest-count failure categories surfaced in Phoenix UI-linked coding sessions.1.1kinstalls2Phoenix Tracingphoenix-tracing is a flat rules-based skill from arize-ai/phoenix that maps OpenInference semantic conventions onto Phoenix instrumentation for LLM apps. It organizes reference files under six semantic prefixes in rules/: span-* for LLM, CHAIN, and TOOL span kinds; setup-* and instrumentation-* for getting started; fundamentals-* and attributes-* for spec reference; plus annotations-* and export-* for advanced trace features. Developers reach for phoenix-tracing when wiring OpenTelemetry spans into chains and agents and aligning attributes with the OpenInference spec and Phoenix OTEL Python API docs.1.1kinstalls3Phoenix Evalsphoenix-evals is an evaluation workflow skill from arize-ai/phoenix that implements axial coding to transform unstructured agent failure notes into actionable, countable categories. The four-step process gathers open coding notes, groups them by shared themes, names actionable categories, and quantifies failure counts per category with YAML taxonomies covering content quality, communication, and context dimensions. Developers building LLM agents reach for phoenix-evals after collecting qualitative eval notes and needing structured failure taxonomies—such as hallucination, tone mismatch, or ignored user context—to drive prioritized prompt, retrieval, and tooling fixes in Phoenix eval workflows.1kinstalls4Agent Browseragent-browser is a browser-automation CLI for AI agents that drives Chrome or Chromium over CDP. Developers use it to navigate pages, fill forms, click buttons, take screenshots, extract data, and run exploratory QA or dogfooding without a Playwright or Puppeteer dependency. It exposes accessibility-tree snapshots with element refs so agents can interact reliably, and ships specialized modes for Electron apps, Slack, and cloud browsers.77installs5Vercel React Best PracticesThis skill is a React and Next.js performance optimization guide maintained by Vercel Engineering. A developer uses it when writing, reviewing, or refactoring React or Next.js code, or optimizing bundle size and load times. It contains 57 rules across 8 categories prioritized by impact, from eliminating request waterfalls and bundle-size cuts to re-render and rendering performance, each with code examples.73installs6Mintlifymintlify is a best-practices skill for building and maintaining documentation sites with the Mintlify platform. Developers use it when creating docs pages, configuring navigation in docs.json, choosing built-in MDX components, or setting up API references. It provides decision tables for navigation patterns and component choices and points to the Mintlify CLI and MCP server.67installs7Phoenix Playwright Testsphoenix-playwright-tests guides writing Playwright end-to-end tests for the Phoenix AI observability platform. Developers use it when creating, updating, or debugging E2E tests that live in app/tests/. It codifies a selector priority order, reusable login and test-credential patterns, common UI interaction recipes (dropdowns, dialogs, tables, tabs), and a centralized timeout policy so tests stay reliable in CI.66installs8Phoenix Pr Screenshotphoenix-pr-screenshot captures screenshots of the Phoenix UI to visually document a feature in a pull request. Developers use it to run the end-to-end flow: build the frontend, start Phoenix with the required env vars, use agent-browser to capture screenshots, upload them to GCS, and update the PR body while preserving existing content. It documents naming conventions and cleanup so PR reviewers can see the feature.35installs9Phoenix Llms Txtphoenix-llms-txt maintains the Phoenix llms.txt documentation index, the machine-readable docs map consumed by AI agents and the px docs fetch CLI. Developers use it when adding, auditing, or reorganizing llms.txt entries following the llmstxt.org spec. It defines inclusion and exclusion rules, targets at least 90% coverage of nav-published pages against the AFDocs threshold, and provides a scripted audit that diffs the docs tree for missing and stale entries.27installs10Phoenix Frontendphoenix-frontend provides frontend development guidelines for the Phoenix AI observability platform. Developers use it when writing, reviewing, or modifying React components, TypeScript, styles, or UI features in the app/ directory. It routes to rule files on components, Relay, accessibility, and logo assets, and requires that significant view state be recreatable from the URL and that visual changes be verified with agent-browser.26installs11Phoenix Release NotesThis skill walks through creating and publishing Phoenix release documentation grounded in real code changes. A developer uses it to identify undocumented releases, analyze commits by reading the changed files, and draft MDX release-note files. It classifies changes into user-facing includes and internal excludes so the notes describe capabilities rather than implementation details.21installs12Phoenix TypescriptThis skill defines TypeScript conventions for all TypeScript in the Phoenix monorepo, spanning the app frontend and the js packages. A developer uses it when writing, reviewing, or refactoring TypeScript such as functions, types, exports, tests, or refactors. It covers self-documenting naming, object-destructured function parameters with JSDoc, type-safety rules, and reuse of shared utilities.19installs13Phoenix Integration Snippetsphoenix-integration-snippets generates onboarding code snippets for Phoenix tracing integrations and wires them into the project onboarding UI. Developers use it to produce install-dependency and implementation code for SDKs like OpenAI, LangChain, and the Vercel AI SDK in Python or TypeScript. It requires testing each snippet verbatim in an isolated environment until traces confirm the integration works before wiring it into the onboarding registry.14installs14Phoenix Githubphoenix-github is a reference for managing GitHub issues, labels, and project boards on the Arize-ai/phoenix repository using the gh CLI. Developers use it to file roadmap epics, triage bugs, apply the correct component and priority labels, and add issues to the Phoenix Roadmap project board. It includes the label taxonomy, roadmap issue templates, and GraphQL recipes for setting project fields.12installs15Phoenix Rest ApiThis skill guides REST API development for the Phoenix server. A developer uses it when adding, modifying, or reviewing endpoints in src/phoenix/server/api/routers/v1/. It provides a pre-commit checklist that regenerates the OpenAPI schema and client types, registers the endpoint in the correct integration-test coverage list by method, and runs the Python linter.11installs16Phoenix Release PleaseThis skill bumps the next release-please version for a Phoenix Python package by opening a PR with a Release-As commit footer. A developer uses it to force a specific version instead of the one release-please would pick from conventional commits. It covers the four managed packages and the squash-merge gotcha where the trailer must live in a commit body to survive.10installs17Phoenix Evals New Metricphoenix-evals-new-metric guides creating a new built-in classification evaluator for Phoenix evals. Developers use it to add an LLM-as-a-judge metric, defining a YAML config as the source of truth that compiles into Python and TypeScript code, then wrapping it in evaluator classes, benchmarking it, and documenting it. It walks a linear pipeline from requirements gathering through YAML, codegen, Python and TS evaluators, builds, and benchmarks.9installs18Phoenix ServerThis skill is a backend development guide for the Phoenix AI observability platform, built on Strawberry GraphQL, async SQLAlchemy, and FastAPI. A developer uses it when adding mutations, types, migrations, or tests in the server and db directories. It enforces hard rules such as putting side effects on Mutation rather than Query to avoid an SSRF vector, plus naming and docstring conventions.9installs19Phoenix Designphoenix-design documents the design-system conventions for the Phoenix frontend, covering layout, dialogs, error display, BEM CSS class naming, and CSS design tokens. Developers use it when building UI, naming CSS classes, creating or consuming tokens, or designing dialog interactions in the app/src directory. It routes to per-task rule files and uses RFC 2119 keywords to state requirements.8installs20Typescript Tooling MigrationThis skill guides migrating or upgrading TypeScript tooling in the Phoenix monorepo across the app frontend and the js packages. A developer uses it when upgrading TypeScript or React, switching tools like ESLint to oxlint or Prettier to oxfmt, or upgrading bundlers. It classifies the work into three migration types and provides a phased workflow covering config inheritance, package scripts, and pre-commit hooks.8installs21Phoenix Docs Gap Auditphoenix-docs-gap-audit finds features in the Phoenix repo that shipped recently without proper documentation. It analyzes recent commits on origin/main (default last 7 days), triages them into audit candidates, and produces a grounded gap report saying what is missing, where it should live, and what the content should say. It checks every doc surface including Mintlify docs, package READMEs, Sphinx and TypeDoc, docstrings, and llms.txt.7installs22Phoenix Typescript Package DocsThis skill maintains the bundled TypeScript package docs that ship inside the Phoenix npm packages for phoenix-client, phoenix-evals, and phoenix-otel. A developer uses it when adding or updating those docs, changing the Mintlify package-doc pages, or aligning bundled docs with actual exports. It defines the canonical MDX source of truth, the sync-and-publish flow, and rules to ground examples in real code.6installs23Phoenix Skills AuditThis skill audits recent changes to Phoenix user-facing surfaces (Python and TypeScript clients, CLI, REST/GraphQL APIs) and patches the three external-facing skills so they match what actually shipped. A developer uses it to catch skill drift, since a stale skill teaches every future agent the wrong API. The output is edits applied to the skill files, defaulting to the last 7 days on origin/main.5installs