skills AGENTS.md

Project instructions for maintaining Langfuse skills and related plugin changes. Langfuse is a platform for observing and evaluating applications that use language models.

In plain words
What is it for?
Use them when adding or improving Langfuse skill use cases, changing skill paths, updating plugin versions, or reviewing pull requests.
Why use it?
They give contributors rules for deciding when a new skill use case is justified and where routing guidance should live, reducing unnecessary documentation and maintenance work.

Instructions file for CodexOpenCode

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add instructions/langfuse/skills/agents-md
Clone the repo
git clone --depth 1 https://github.com/langfuse/skills

Made for: Codex, OpenCode.

Per session 1,518 This file is loaded in full into every session.
When invoked 1,518 The same file — it is already loaded in full.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.01518 $0.01518
Opus 5 $0.00759 $0.00759
Sonnet 5 $0.00304 $0.00304
Haiku 4.5 $0.00152 $0.00152

Measured 2d ago against content hash 4abe8d3026c5, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

skills AGENTS.md scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

AGENTS.md · 69 lines

How it starts

The opening of the file, as written. The whole thing — 69 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Agent Instructions

Adding or Improving a Skill Use Case

FOLLOW THESE INSTRUCTIONS RELIGIOUSLY. After every edit you make, come back to these principles and judge critically whether you adhered to them. Fix your edits if not.

  • Only add a use case if it beats the docs. If an agent can already serve the user by fetching the Langfuse docs, add nothing. Reserve new content for where docs fall short and the agent needs extra context. Every addition is maintenance surface and dilutes the skill.

  • You should almost never touch the top-level frontmatter description in SKILL.md. It only controls whether the skill is invoked, and a user asking about a use case already mentions Langfuse or evaluation — which triggers it. Keep it short; in-skill routing handles the rest.

  • Put "when to use" guidance in exactly two places: exactly one one-line entry per reference in the ## Use case specific references list in SKILL.md, and the description in the reference file's frontmatter. Nowhere else — no prose routing section, no "when to use" section in the reference body. A reference body is read only after the agent already chose to open it, so routing text there is dead weight.

  • Every reference file's frontmatter must declare a metadata.required_access list — the kinds of access an agent needs to execute that reference. Use only these tokens, and reuse them consistently across files:

    • CODEBASE — reads or edits the user's source code
    • LANGFUSE_PROJECT_INTERFACE — reaches the Langfuse project via CLI / API / MCP commands
    • LANGFUSE_PROJECT_SCRIPT — runs SDK code that connects to the Langfuse backend (needs network)
    • GITHUB — operates on GitHub via the gh CLI
  • Every line must earn its place. Add only what's useful or what an agent couldn't infer on its own. Cut filler, restatements, self-explanatory steps, and anything the agent will already have from the relevant docs or task context. Keep new use-case references at 100 lines or fewer, including frontmatter; this is a maximum, if you exceed this length, you are likely producing slop.

Read the full file on GitHub · 69 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 2d ago First seen · 69 lines · 1,518 tokens per session scan A 4abe8d3026c5

Subscribe to this mod's changes

skills AGENTS.md is an instructions file published in the GitHub repository langfuse/skills (266 stars, last pushed 4d ago), licensed MIT. It adds 1,518 tokens to every session, about $0.0076 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.