braintrust-claude-plugin AGENTS.md

Repository instructions for a Claude Code plugin marketplace, a project that distributes installable coding-agent extensions. They describe the repository layout, its plugins, and key terms.

In plain words
What is it for?
Use them when working on the Braintrust evaluation and logging plugin, the Claude Code session-tracing plugin, or their evaluation suite.
Why use it?
They give an agent the project-specific context needed to edit and maintain the marketplace consistently.

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/braintrustdata/braintrust-claude-plugin/agents-md
Clone the repo
git clone --depth 1 https://github.com/braintrustdata/braintrust-claude-plugin

Made for: Codex, OpenCode.

Per session 357 This file is loaded in full into every session.
When invoked 357 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 $0.00357 $0.00357
Opus 5 $0.00179 $0.00179
Sonnet 5 $0.00071 $0.00071
Haiku 4.5 $0.00036 $0.00036

Measured 2d ago against content hash 6f61da683cca, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

braintrust-claude-plugin 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 · 37 lines

What it actually says

Agent guidelines

About this repository

This is the Braintrust Claude Code plugin marketplace - a repository that distributes Claude Code plugins for Braintrust integration.

Structure

claude-plugin/
├── .claude-plugin/
│   └── marketplace.json      # Marketplace catalog (lists available plugins)
├── plugins/
│   ├── braintrust/           # Plugin: Braintrust evaluation & logging
│   └── trace-claude-code/    # Plugin: Session tracing to Braintrust
└── evals/                    # Evaluation suite for testing the plugins

Plugins

Plugin Description
braintrust Enables AI agents to use Braintrust for LLM evaluation, logging, and observability. Includes MCP server config and the troubleshoot-braintrust-mcp skill.
trace-claude-code Forwards Claude Code lifecycle hooks to bt trace hook --source claude-code; the shared daemon builds and delivers traces.

Terminology

  • Marketplace: A repository with a marketplace.json that catalogs multiple plugins for distribution
  • Plugin: An installable unit with its own .claude-plugin/plugin.json manifest
  • Skill: A capability within a plugin (e.g., troubleshoot-braintrust-mcp is a skill in the braintrust plugin)

Style conventions

  • Use sentence case for all text (capitalize first word only, except for proper nouns and code references)
  • Keep criteria concise and specific
  • Reference exact function/method names with proper casing (e.g., init_dataset(), Eval(), Factuality)
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 · 37 lines · 357 tokens per session scan A 6f61da683cca

Subscribe to this mod's changes

braintrust-claude-plugin AGENTS.md is an instructions file published in the GitHub repository braintrustdata/braintrust-claude-plugin (20 stars, last pushed 20d ago), licensed MIT. It adds 357 tokens to every session, about $0.0018 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.