What is agentxp CLAUDE.md?

A project-specific instruction set for AgentXP, a system for designing and analysing controlled experiments. It defines which skills handle experiment design, analysis, audits, data connections, and readouts.

Use it to decide how AgentXP should pre-register experiments, analyse sealed briefs, inspect experiment history, connect data, and produce readouts.

It keeps experiment registration separate from analysis and clarifies which workflow applies to each request.

How ai-analyst-lab/agentxp configures Claude Code

Reuse

Borrowing it

Nothing to install: this file belongs to ai-analyst-lab/agentxp. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.

Copy the file

curl -O https://raw.githubusercontent.com/ai-analyst-lab/agentxp/main/CLAUDE.md

Clone the repo

git clone --depth 1 https://github.com/ai-analyst-lab/agentxp

Made for: Claude Code.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

README.md

[![agentmods](https://agentmods.dev/badge/instructions/ai-analyst-lab/agentxp/claude-md/github.svg)](https://agentmods.dev/instructions/ai-analyst-lab/agentxp/claude-md)

Your own site

<a href="https://agentmods.dev/instructions/ai-analyst-lab/agentxp/claude-md"><img src="https://agentmods.dev/badge/instructions/ai-analyst-lab/agentxp/claude-md/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

Your own site · 80×15

<a href="https://agentmods.dev/instructions/ai-analyst-lab/agentxp/claude-md"><img src="https://agentmods.dev/badge/instructions/ai-analyst-lab/agentxp/claude-md.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 3,971 This file is loaded in full into every session.

When invoked 3,971 The same file — it is already loaded in full.

Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.

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.

Model	Per session	Once invoked
Fable 5.1	$0.03971	$0.03971
Opus 5.5	$0.01588	$0.01588
Sonnet 5.5	$0.00794	$0.00794
Haiku 4.5	$0.00397	$0.00397

Measured 2mo ago against content hash 3db44076891c, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

agentxp CLAUDE.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 2mo 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.

CLAUDE.md · 188 lines

How it starts

The opening of the file, as written. The whole thing — 188 lines — stays where its author put it; the contents beside it link to each section on GitHub.

CLAUDE.md — AgentXP

1. Identity

AgentXP is a single-user system for the design and analysis of controlled experiments, opened inside Claude Code and driven by an orchestrator agent (you) who follows the discipline below. Two verb skills are an architectural wall: design pre-registers, analyze runs against a sealed brief. You never bridge them mid-session. Three other skills (audit, readouts, connect-data) are tools, not verbs — they don't move an experiment through its lifecycle, they help you look at or set up the surrounding state.

2. Skills registry

Skill	Kind	Path	Apply when	Terminal artifact
`design`	verb	`.claude/skills/design/`	user wants to pre-register an experiment	`brief.sealed.yaml`
`analyze`	verb	`.claude/skills/analyze/`	user wants to analyze a sealed brief	verdict share-tail (`readouts/verdict/<persona>/<persona>.{html,pdf}` + catalog row)
`audit`	tool	`.claude/skills/audit/`	user asks what happened in an experiment	text/HTML timeline
`readouts`	tool	`.claude/skills/readouts/`	user wants to walk the renders catalog	catalog rows or `readouts/index.html`
`connect-data`	tool	`.claude/skills/connect-data/`	user needs to wire a warehouse	`~/.agentxp/credentials/<dialect>/<profile>.yaml`

Resuming an in-flight experiment is not a slash command (/resume is reserved by Claude Code). The first-turn behavior (§8) lists in-flight experiments via agentxp.workflows.resume.list_in_flight; the user routes to /design --exp-id <existing> (pre-seal) or /analyze --brief <path> (post-seal).

3. Specialists

Six roles total. The orchestrator's prompt is this file. The five specialist prompts live at agents/<role>.md with a machine-readable CONTRACT block at the top of each. Full roster + DAG in agents/INDEX.md; generated DAG in agents/registry.yaml.

The five specialists don't all fire on the same cadence:

Per-experiment, always dispatched: designer (3 tasks: hypothesis, brief, data plan), critic (≥4 times per experiment), analyst_narrator (once at verdict).
Setup-only, dispatched when needed: understander fires only when semantic_models/ or metrics/ are missing or stale. In a healthy project it fires zero times per experiment.
Reserved (contract exists, not yet wired): sql_specialist has a complete agent prompt + SqlSpecialistBundle schema, but no skill currently dispatches it. The orchestrator writes SQL directly through probe_data. The role is held against a future "specialist drafts the SQL" workflow; today it's a roster placeholder.

Read the full file on GitHub · 188 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.

2mo ago First seen · 188 lines · 3,971 tokens per session scan A 3db44076891c

Subscribe to this mod's changes

agentxp CLAUDE.md is an instructions file published in the GitHub repository ai-analyst-lab/agentxp (12 stars, last pushed 2mo ago), licensed Apache-2.0. It adds 3,971 tokens to every session, about $0.0159 per session on Opus 5.5. 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.

Planning Scientific

agentxp CLAUDE.md next to 6 instructions from other repositories
Mod	Type	Stars	Last push	Licence	Scan	Tokens	Works with
agentxp CLAUDE.mdai-analyst-lab/agentxp	Instructions file	12	2mo ago	Apache-2.0	A	3,971	Claude Code
vscode buildNext.instructions.mdmicrosoft/vscode	Instructions file	195k	yesterday	MIT	A	9,719	GitHub Copilot
next.js AGENTS.mdvercel/next.js	Instructions file	143k	yesterday	MIT	A	7,461	Codex, OpenCode
langchain AGENTS.mdlangchain-ai/langchain	Instructions file	149k	yesterday	MIT	A	4,680	Codex, OpenCode
vscode oss-third-party-notices.instructions.mdmicrosoft/vscode	Instructions file	195k	yesterday	MIT	A	5,516	GitHub Copilot
vscode copilot-instructions.mdmicrosoft/vscode	Instructions file	195k	yesterday	MIT	A	3,093	GitHub Copilot
deepseek-harness AGENTS.mddeepseek-ai/deepseek-harness	Instructions file	274k	2d ago	MIT	A	4,313	Codex, OpenCode

agentxp: Instructions file for Claude Code