What is arena-agent AGENTS.md?

Maintainer instructions for arena-agent, a modular Python project for running and managing AI-agent features. They set architectural limits and rules for where runtime logic, wrappers, installers, and platform-specific code belong.

Use them when adding or reorganising arena-agent modules, changing command-line wrappers or installers, handling network providers, or running the architecture and modularity tests.

They reduce the risk of creating oversized modules, breaking platform support, or hiding installation failures.

How IvanSkainet/arena-agent configures Codex and OpenCode

Reuse

Borrowing it

Nothing to install: this file belongs to IvanSkainet/arena-agent. 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/IvanSkainet/arena-agent/master/AGENTS.md

Clone the repo

git clone --depth 1 https://github.com/IvanSkainet/arena-agent

Made for: Codex, OpenCode.

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/ivanskainet/arena-agent/agents-md.svg)](https://agentmods.dev/instructions/ivanskainet/arena-agent/agents-md)

Your own site

<a href="https://agentmods.dev/instructions/ivanskainet/arena-agent/agents-md"><img src="https://agentmods.dev/badge/instructions/ivanskainet/arena-agent/agents-md.svg" alt="Measured on agentmods" height="20"></a>

Per session 7,880 This file is loaded in full into every session.

When invoked 7,880 The same file — it is already loaded in full.

Security scan A 2 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	$0.07880	$0.07880
Opus 5	$0.03940	$0.03940
Sonnet 5	$0.01576	$0.01576
Haiku 4.5	$0.00788	$0.00788

Measured 2d ago against content hash fab98d9c232e, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-07, from the pricing page.

Security

Grade A, and why

arena-agent AGENTS.md scanned grade A with 2 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 2d 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.

Makes network callslowCapability

Not a fault in itself. Listed so you know the mod talks to something, and to what.

curl -sH "Authorization: Bearer $(cat token.txt)" \

Runs shell commandslowCapability

Expected in a hook, worth knowing in a rule or an instructions file.

- **Never use `os.system()`** — always argv-form

AGENTS.md · 584 lines

How it starts

The opening of the file, as written. The whole thing — 584 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Arena Agent Codebase Guide for AI Maintainers

This repository is intentionally modular. Do not add new runtime logic to thin compatibility entrypoints or large catch-all files.

Hard rules

Architecture

Keep unified_bridge.py a thin compatibility/CLI entrypoint.
Keep wrapper scripts in scripts/ and bin/ thin; real logic belongs under arena/.
Product files must stay under the modularity limit enforced by tests/test_project_modularity.py (currently 1600 lines). Runtime modules under arena/ have an additional limit in tests/test_architecture_boundaries.py (currently 600 lines). Readable code beats squeezed code — if a file is close to the limit, split it by responsibility instead of collapsing whitespace.
Do not import unified_bridge.py from arena/* modules.
Every new module in arena/admin/ (tunnels/network providers) must be cross-platform: platform.system() branches, no Linux-only assumptions, never invoke sudo directly.
Every user-facing installer path (install.sh, install.bat) must verify after installing dependencies — never trust a silent pip install ... 2>/dev/null || true.
Never store per-release scratch notes in the repository; use /tmp/ when driving gh release create --notes-file.

Task Board Discipline & Spec-Kit Protocol (T0..Tn) (v4.169.39+)

Every action belongs to a tracked task. Any newly discovered bug, refactoring, or feature must be recorded in docs/TASK_BOARD.md with an identifier T0..Tn. Never execute ad-hoc or unrecorded side tasks.
Strict Definition of Done (DoD):
1. Root-cause fix: resolve the true invariant failure, never suppress or silence.
2. Deterministic Parity Suite: write isolated unit tests without unmocked sockets or clocks.
3. 0 Mutation Survivors: verify with mutmut and pin the ratchet in scripts/mutation_baseline.json.
4. Mandatory Bilateral Sabotage: prove that breaking the fix fails tests, and healthy code is 100% green.
5. Preflight Pass: python scripts/preflight.py must be completely green before release.
Sandbox Escape via Bridge: Never download heavy engines (like Godot), 3D assets, or datasets into the 128 MB agent workspace /home/user. Route heavy workloads, game production, and persistent storage to the host machine via Skainet Bridge (skills/arena-bridge/SKILL.md and docs/GODOT_INTEGRATION.md).
Serena Continuity: To prevent context loss across session truncations, read docs/TASK_BOARD.md and run python scripts/serena_reminder.py upon session bootstrap.

Read the full file on GitHub · 584 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.

2d ago Changed · +30 lines · +404 tokens per session fab98d9c232e
7d ago First seen · 554 lines · 7,476 tokens per session scan A 2676c3f0ad8e

Subscribe to this mod's changes

arena-agent AGENTS.md is an instructions file published in the GitHub repository IvanSkainet/arena-agent (5 stars, last pushed yesterday), licensed MIT. It adds 7,880 tokens to every session, about $0.0394 per session on Opus 5. A static security scan graded it A with 2 findings (makes network calls, runs shell commands). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

agentic-ai ai-agent ai-agents automation browser-automation claude

agent-compass pr-workflow.instructions.md

Instructions for me-cedric/agent-compass: For PR creation, default base branch is develop. Assign the PR to yourself with --assignee @me. If reviewers are not specified, ask for at least one and list likely repo contributors. Use only labels that exist in gh label list.

me-cedric/agent-compass · 96 tokens

codex AGENTS.md

AGENTS.md instructions for openai/codex, covering rust/codex-rs, the codex-core crate, code review rules, crate api surface and model visible context.

openai/codex · 5,182 tokens

vscode buildNext.instructions.md

Working notes and architecture documentation for the new esbuild-based build system in build/next. Use when making changes to the new build pipeline (transpile/bundle commands, NLS plugin, source-map handling, resource copying, or self-hosting watch tasks).

microsoft/vscode · 6,785 tokens

arena-agent: Instructions file for Codex