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.
npx agentmods add instructions/neat-technologies/neat/claude-mdgit clone --depth 1 https://github.com/neat-technologies/neatWrote 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.
[](https://agentmods.dev/instructions/neat-technologies/neat/claude-md)<a href="https://agentmods.dev/instructions/neat-technologies/neat/claude-md"><img src="https://agentmods.dev/badge/instructions/neat-technologies/neat/claude-md.svg" alt="Measured on agentmods" height="20"></a>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 | $0.00909 | $0.00909 |
| Opus 5 | $0.00454 | $0.00454 |
| Sonnet 5 | $0.00182 | $0.00182 |
| Haiku 4.5 | $0.00091 | $0.00091 |
Grade A, and why
neat 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 6d 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.
How it starts
The opening of the file, as written. The whole thing — 51 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CLAUDE.md
Agent guide for the NEAT repo. Read this first if you're a fresh Claude session or a human picking this up cold.
Binding rules
@docs/contracts.md is the index, auto-loaded with this file. Per-topic contracts live under docs/contracts/ and are surfaced automatically when you edit a governed file (PreToolUse hook at docs/contracts/_hook.sh, wired in .claude/settings.json). If you write code that conflicts with a contract, stop. The conflict is the bug.
What NEAT is
NEAT solves the AI coding context problem. It keeps a live semantic graph of a software system — code, infrastructure, and runtime behavior fused into one model, queryable over MCP and a REST API — so an AI agent has accurate, full-stack context: enough to build whole feature sets and debug autonomously, instead of grepping files and guessing. The graph carries provenance on every edge: EXTRACTED from source, OBSERVED from OpenTelemetry, INFERRED where the trace stitcher bridges gaps, STALE when runtime stops speaking — so the agent knows how much to trust each claim. Divergence between declared intent and observed reality is one of the questions that fusion makes answerable; root-cause, blast-radius, and policy checks are others. The graph is the product; those queries are features of it, not the point.
The extraction pipeline reads static code via tree-sitter (JavaScript, TypeScript, Python) and ingests live OTel spans to build and maintain the graph.
What success looks like
NEAT earns its keep when an agent, using the graph as its eyes, builds or debugs a real system more autonomously and accurately than the same agent without it — on a codebase NEAT was not engineered against. The OBSERVED layer carries the load: fusing runtime with static lets the agent see what the system actually does, not only what it declares. Static analysis alone is what other tools already do. A real bug surfaced along the way — say, through a divergence query — is evidence the model works; agent autonomy is the goal.
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.
- 6d ago First seen · 51 lines · 909 tokens per session scan A 8209c5442c0f
neat CLAUDE.md is an instructions file published in the GitHub repository neat-technologies/neat (18 stars, last pushed 4d ago), licensed Apache-2.0. It adds 909 tokens to every session, about $0.0045 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.
Other instructions, from other repositories
mcp-adr-analysis-server copilot-instructions.md
Copilot instructions for tosin2013/mcp-adr-analysis-server, covering mcp adr analysis server - ai coding agent instructions, project overview, critical technical conventions, esm-only module system and typescript configuration.
chiplab copilot-instructions.md
Instructions for veecle/chiplab: Agent instructions for this repository live in AGENTS.md at the repo root. Please read that file for repository conventions, available workflows, and guidelines.
symphony AGENTS.md
Instructions for broomva/symphony, covering agents.md - symphony, repository purpose, architecture, key design decisions and gathering context from the knowledge graph.
cosmos-os CLAUDE.md
Instructions for ludeo-labs/cosmos-os, covering the cosmos — notes for claude code, commands, layout and invariants.
ai-software-architect AGENTS.md
AGENTS.md instructions for codenamev/ai-software-architect, covering agents.md - ai software architect framework, project overview, framework development setup, repository structure and installation for framework development.
agentic-bim-team CLAUDE.md
Instructions for louistrue/agentic-bim-team, covering bim agent team — swiss residential construction, ifc analysis & creation (ifc-lite cli), quick reference, ifc create element types and supported --from-json fields per type (examples).