Borrowing it
Nothing to install: this file belongs to belt-sh/cli. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/belt-sh/cli/main/AGENTS.mdgit clone --depth 1 https://github.com/belt-sh/cliWrote 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/belt-sh/cli/agents-md)<a href="https://agentmods.dev/instructions/belt-sh/cli/agents-md"><img src="https://agentmods.dev/badge/instructions/belt-sh/cli/agents-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.
<a href="https://agentmods.dev/instructions/belt-sh/cli/agents-md"><img src="https://agentmods.dev/badge/instructions/belt-sh/cli/agents-md.svg" alt="Reviewed on agentmods" width="80" 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 | $0.00476 | $0.00476 |
| Opus 5.5 | $0.00190 | $0.00190 |
| Sonnet 5.5 | $0.00095 | $0.00095 |
| Haiku 4.5 | $0.00048 | $0.00048 |
Grade A, and why
cli 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 1mo 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.
What it actually says
agents working with belt
this file provides context for any ai agent (claude, cursor, copilot, etc.) working in a project where belt is available.
what belt is
belt is a cli that gives agents access to:
- hundreds of ai apps — image generation (gpt-image, reve, pruna), video (veo, seedance), audio (tts, music), llms (claude, gemini), search (tavily, exa), 3d, and more
- knowledge system — persistent memory that survives across sessions. save observations, concepts, skills, references
- skill registry — reusable agent workflows. search, install, publish
- mcp connectors — connect to any mcp-compatible service through the platform
when to use belt
- generating media:
belt app search "image"→belt app run openai/gpt-image-2 --input '{"prompt": "..."}' - remembering something:
belt know create - --name finding --type observation <<< "learned X" - looking for a workflow:
belt skill search "deploy"orbelt suggest "how to do X" - connecting to services:
belt mcp connect slack→belt mcp run slack send_message - web search / research:
belt app run tavily/search --input '{"query": "..."}'
output format
- tty mode: human-readable, colored output
- piped/
--json: structured json, safe for parsing - always use
--jsonwhen you need to process output programmatically
conventions
- app inputs are always json (
--inputflag or file path) - knowledge types:
skill,concept,observation,reference,preference - skill names follow
namespace/nameformat - apps follow
owner/app-nameorowner/app-name@versionformat
error handling
belt exits non-zero on failure with a descriptive error message. common issues:
- not authenticated →
belt login - app not found → check spelling, try
belt app search - invalid input → use
belt app sampleto generate valid input schema
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.
- 1mo ago Changed · +1 tokens per session 438adc39fb5f
- 2mo ago First seen · 41 lines · 475 tokens per session scan A 4760ea04b1a6
cli AGENTS.md is an instructions file published in the GitHub repository belt-sh/cli (7 stars, last pushed 1mo ago), licensed MIT. It adds 476 tokens to every session, about $0.0019 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 2026-08-31.
cli AGENTS.md compared with instructions from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| cli AGENTS.mdbelt-sh/cli | Instructions file | 7 | 1mo ago | MIT | A | 476 | Codex, OpenCode |
| vscode buildNext.instructions.mdmicrosoft/vscode | Instructions file | 195k | today | 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 | today | MIT | A | 4,680 | Codex, OpenCode |
| vscode oss-third-party-notices.instructions.mdmicrosoft/vscode | Instructions file | 195k | today | MIT | A | 5,516 | GitHub Copilot |
| vscode copilot-instructions.mdmicrosoft/vscode | Instructions file | 195k | today | MIT | A | 3,093 | GitHub Copilot |
| deepseek-harness AGENTS.mddeepseek-ai/deepseek-harness | Instructions file | 275k | 2d ago | MIT | A | 4,313 | Codex, OpenCode |