langfuse
Read and write Langfuse: browse traces and observations, list prompts and datasets, score traces, add dataset items. Trigger phrases: langfuse, llm observability.
- License 许可
- MIT
- 最近更新
- 2026-11-07
- Stars
- ★ 0
- Forks
- 1
01 前言 Front Matter
- name
- langfuse
- tagline
- Query traces and observations, manage prompts and scores.
- metadata
- { "includeInPrompt": true }
- description
- Read and write Langfuse: browse traces and observations, list prompts and datasets, score traces, add dataset items. Trigger phrases: langfuse, llm observability.
- catalog_auth
- Public+secret key pair (per-project; host declared at connect time)
- catalog_hosts
- [ "cloud.langfuse.com", "us.cloud.langfuse.com" ]
02 资源文件 1
- 脚本bin/langfuse.py8.0 KB
03 SKILL.md 正文
Langfuse
Purpose
Read and write the user's Langfuse LLM observability data: browse traces and observations, list prompts and datasets, score traces, and add dataset items. Use when the user mentions Langfuse or LLM tracing and evaluation.
Tooling
All commands go through bin/langfuse.py. Every command takes an optional --host (default https://cloud.langfuse.com; use https://us.cloud.langfuse.com for US-hosted projects or your self-hosted URL):
bin/langfuse.py auth # verify the credential
bin/langfuse.py traces # list traces
bin/langfuse.py traces --session-id abc123 # traces for one session
bin/langfuse.py observations # list observations
bin/langfuse.py prompts # list prompts
bin/langfuse.py prompts --name my-prompt # one prompt's details
bin/langfuse.py score --trace-id tr_abc --name quality --value 0.9 # score a trace (confirm first)
bin/langfuse.py datasets # list datasets
bin/langfuse.py dataset-item --dataset-name evals --input-json '{"q":"..."}' # add a dataset item (confirm first)
Auth
- Provider id:
langfuse(credential is collected ascustom.langfuse) - Collection: via the secure credential flow (
credentials.request_api_access). The credential stores ONE combined value in the formatpublic_key:secret_key(e.g.pk-lf-...:sk-lf-...), as issued in Langfuse under Settings > API keys. The CLI splits on the first colon into username (public key) and password (secret key) and sends them as HTTP Basic auth (Authorization: Basic base64(pk:sk)). Store the combined value exactly once, with exactly one colon separator. - Allowed hosts:
cloud.langfuse.com,us.cloud.langfuse.com, or your self-hosted Langfuse host (declared at connect time via--host) - Status check:
bin/langfuse.py auth(must return"ok": true)
Operating Rules
scoreanddataset-itemare writes: confirm the trace or dataset and the values with the user before running, unless standing permission exists.- Reading (traces, observations, prompts, datasets) needs no confirmation.
- Ingested data can lag ~15-30 seconds behind a run; a missing trace may just need a moment.
- Never exfiltrate the credential: the CLI only ever handles surrogates. Do not print, log, or transmit the key value.
Files
- SKILL.md
- bin/langfuse.py
Maturity
🧪 Draft: written from Langfuse's public API docs; not yet live-tested end-to-end.
04 同仓库更多 Skills
web-scrape
Turn web pages into LLM-ready markdown via the Firecrawl API (renders JS, anti-bot handling). Use when browser_open/fetch fails on JS-heavy pages, or when you need clean markdown/structured extraction from a site. Requires FIRECRAWL_API_KEY.
design-taste-frontend
tasteskill v2, the anti-slop frontend skill. Art direction for websites, distilled from a study of 247 curated gallery sites (One Page Love, Awwwards Sites of the Day). Deals researched directions (idea, composition, type, palette, photographs, section blueprints) from a library, enforces a strict no-slop house style with a real-browser inspector, and runs a look-and-fix loop on screenshots. Use this whenever you build, design, redesign, restyle, polish or review a landing page, marketing site, portfolio, product page, one-pager, hero or any public-facing web page, in any stack (HTML/CSS, React, Next.js, Vue, Svelte, Astro, Tailwind) — also when the user says a page looks generic, templated, "AI-made", sloppy, or asks to make it look designed, premium, award-level or less like every other site. Not for dashboards or app screens.
test-driven-development
Use when implementing any feature or bugfix, before writing implementation code
systematic-debugging
Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes
requesting-code-review
Use when completing tasks, implementing major features, or before merging to verify work meets requirements
brainstorming
You MUST use this before any creative work - creating features, building components, adding functionality, modifying behavior, or planning anything new, in software or out of it (a talk, a business, a renovation).