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/kyosukeishizu1008/berryscode/agents-mdgit clone --depth 1 https://github.com/KyosukeIshizu1008/berryscodeWrote 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/kyosukeishizu1008/berryscode/agents-md)<a href="https://agentmods.dev/instructions/kyosukeishizu1008/berryscode/agents-md"><img src="https://agentmods.dev/badge/instructions/kyosukeishizu1008/berryscode/agents-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 | $0.01886 | $0.01886 |
| Opus 5 | $0.00943 | $0.00943 |
| Sonnet 5 | $0.00377 | $0.00377 |
| Haiku 4.5 | $0.00189 | $0.00189 |
Grade A, and why
berryscode 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 5d 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 — 87 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md
Guidance for Codex working in this repo. Keep this file short — codebase facts belong in code, README, or commit messages, not here. Only document non-obvious gotchas and conventions an agent can't infer in 30 seconds.
What this is
BerryCode is a native IDE built specifically for the Bevy game engine. Same stack as the games it edits: Bevy 0.18 + bevy_egui 0.39 + egui 0.33 + WGPU. Two binaries ship from the same berrycode crate:
berrycode(src/main.rs) — the production binaryberrycode-egui(src/bin/berrycode-egui.rs) — a thinner harness used in dev / testing
Both wire up BerryCodePlugin from src/bevy_plugin.rs. Add new app-level systems to the plugin, not to either bin. Otherwise berrycode-egui silently misses behavior.
Build / run
cargo run --bin berrycode # debug
cargo build --release --bin berrycode # release
cargo check -p berrycode # fast iteration when iterating UI code
The dev build is large; cargo check is the right loop for editing egui UI. The app uses a PID lockfile at ~/Library/Caches/berrycode.lock (macOS) — if the binary crashes mid-run, delete it before relaunching.
Operating rules (read first)
These exist to stop the "止まった?" feedback loop. Follow them by default; only break them if the user explicitly asks.
Never let the user lose the thread
- Always announce intent before the first tool call. When the user asks for something, the first thing in the reply is one line stating what you're about to do — not silence followed by tool calls. The user must never see "thinking…" with no follow-up text.
- Every assistant turn must end with one of three things, never silence: (a) a concrete result statement, (b) a single specific question that unblocks you, or (c) "next: " when you're handing off to a tool that fires asynchronously. Never end with "確認待ちです" alone — pair it with the specific thing you need.
- Don't send "ビルド待ち" / "monitoring" filler. If you've armed a Monitor, the next message should come from the Monitor event, not from you. Filler messages create the appearance of stalls.
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.
- 5d ago First seen · 87 lines · 1,886 tokens per session scan A 38642f3ca4df
berryscode AGENTS.md is an instructions file published in the GitHub repository KyosukeIshizu1008/berryscode (144 stars, last pushed 3mo ago), licensed MIT. It adds 1,886 tokens to every session, about $0.0094 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
comedot AGENTS.md
AGENTS.md instructions for InvadingOctopus/comedot, covering agents.md, guidelines for the comedot project and repository, project overview, folder structure & module organization and subsystems.
bevy-agent-feedback-plugin AGENTS.md
AGENTS.md instructions for drsh4dow/bevy-agent-feedback-plugin, covering context, general rules, code design, ci gate and communication rules.
Legends-Of-Heroes AGENTS.md
Instructions for FlameskyDexive/Legends-Of-Heroes, covering agents.md, 最小入口规则, 两套技能体系分工, unity 交互入口优先级 and aibridge bootstrap.
everything-game-dev-code copilot-instructions.md
Copilot instructions for MRCalderon3D/everything-game-dev-code, covering github copilot repository instructions, repository intent, how to navigate the repository, working rules and repository maintenance.
codex-game-studio AGENTS.md
AGENTS.md instructions for merlinhu1/codex-game-studio, covering codex game studio template, project goal, engine, commands and context bootstrap.
GameDev-MCP-Server CLAUDE.md
Instructions for IvanMurzak/GameDev-MCP-Server, covering claude.md, what this is, build / run and versioning / releases.