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/buzz-language/buzz/agents-mdgit clone --depth 1 https://github.com/buzz-language/buzzWhat 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 | $0.01851 | $0.01851 |
| Opus 5 | $0.00925 | $0.00925 |
| Sonnet 5 | $0.00370 | $0.00370 |
| Haiku 4.5 | $0.00185 | $0.00185 |
Grade A, and why
buzz 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 yesterday.
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 — 174 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md
Guidance for coding agents working in this repository.
Project
buzz is a small/lightweight statically typed scripting language written in Zig.
Main buzz implementation subsystems:
- Scanner:
src/Scanner.zig - Parser and AST:
src/Parser.zig,src/Ast.zig - Code generation and bytecode:
src/Codegen.zig,src/Chunk.zig - Runtime values and objects:
src/value.zig,src/obj.zig - VM:
src/vm.zig - GC:
src/GC.zig - JIT:
src/Jit.zig,src/mir.zig - FFI:
src/FFI.zig,src/lib/buzz_ffi.zig,src/lib/ffi.buzz
Tooling subsystems:
- Debugger:
src/Debugger.zig - LSP:
src/lsp.zig - Formatter:
src/renderer.zig,src/tests/fmt.zig - REPL:
src/repl.zig,src/repl.html,src/wasm_repl.zig
FFI, Debugger, and LSP are immature. Treat changes there as higher risk.
Working Tree Policy
- Never commit changes.
- Before editing, check the working tree.
- If the working tree is clean, make the requested changes normally.
- If the working tree only has untracked files, edits are allowed. Do not overwrite or delete unrelated untracked files.
- If the working tree has tracked modifications or staged changes that were not initiated by the current agent/session, continue investigation read-only, then show the diff you would have applied and say that no files were modified because the working copy was not clean.
- If tracked modifications were initiated by the current agent/session for the active task, the agent may continue editing those files and related files needed to finish the same task.
- Do not overwrite or revert user changes.
Agents may delete files they generated during their own work. Do not delete files that were already tracked by git.
Do Not Touch
- Never edit
vendors/. - Do not edit generated or local build artifacts, including
zig-cache/,zig-out/,dist/,build/,node_modules/, object files, dylibs, tarballs, and platform package directories. - Only edit
package.json,package-lock.json, or WASM/frontend files when the task is specifically web/WASM-related. - Do not download or install dependencies.
- If submodules are missing,
git submodule update --initis acceptable. - Do not run
scripts/perf_compare_commits.py.
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.
- yesterday First seen · 174 lines · 1,851 tokens per session scan A 74a32a60c2e7
buzz AGENTS.md is an instructions file published in the GitHub repository buzz-language/buzz (1,529 stars, last pushed 5d ago), licensed MIT. It adds 1,851 tokens to every session, about $0.0093 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
nullclaw AGENTS.md
Instructions for nullclaw/nullclaw, covering agents.md — nullclaw agent engineering protocol, 1) project snapshot (read first), 3) engineering principles (normative), 3.1 kiss and 3.2 yagni.
nullclaw CLAUDE.md
Instructions for nullclaw/nullclaw, covering claude.md, mandatory reference, build & test commands, requires exactly zig 0.16.0 (verify: zig version) and build flags.
pipelex AGENTS.md
Instructions for Pipelex/pipelex, covering pipelex coding rules, codex cloud commands, linting, keyword-only arguments check and cleaning derived files.
pipelex CLAUDE.md
Instructions for Pipelex/pipelex, covering pipelex coding rules, commands, linting, keyword-only arguments check and cleaning derived files.
zvm CLAUDE.md
Instructions for tristanisham/zvm, covering claude.md, project overview, build & test commands, architecture and packaging & generated files.
zvm AGENTS.md
Instructions for tristanisham/zvm, covering repository guidelines, project structure & module organization, build, test, and development commands, coding style & naming conventions and testing guidelines.