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/neelprime/arbiterx/agents-mdgit clone --depth 1 https://github.com/NeelPrime/arbiterxWhat 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 | $0.00345 | $0.00345 |
| Opus 5 | $0.00172 | $0.00172 |
| Sonnet 5 | $0.00069 | $0.00069 |
| Haiku 4.5 | $0.00034 | $0.00034 |
Grade A, and why
arbiterx 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 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.
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
ArbiterX — Engineering Discipline for AI Code Generation
Before Writing Code
Run this ladder. Stop at the first "yes":
- Does this need to exist? → No → skip (YAGNI)
- Already in this codebase? → Reuse it
- Stdlib does it? → Use stdlib
- Installed dep does it? → Use it
- One-liner? → One line, no abstraction
- Only then → write the minimum that works
The 10 Rules (Always Enforce)
- Type all function signatures (params + return)
- Handle every error path (no bare except, no swallowed errors)
- Close every resource (use
with, never bare open) - Name all constants (no magic numbers)
- One function = one job (max 30 lines)
- Validate inputs at function entry
- No dead code (no commented code, no unused imports)
- Finish everything (no TODO/FIXME — implement or skip)
- Prefer immutable (frozen dataclasses, tuples)
- Every line earns its place (delete what doesn't change behavior)
Never Generate
- Hardcoded secrets (API keys, passwords, tokens)
- SQL with string formatting (use parameterized queries)
- eval() or exec()
- subprocess with shell=True
- Network calls without timeouts
- File ops without error handling
After Writing
- Can any line be deleted without breaking behavior? → Delete it
- Is there a simpler stdlib approach? → Use it
- Would this pass review by the strictest engineer you know? → If not, fix it
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 First seen · 41 lines · 345 tokens per session scan A 520c4f5a6ff9
arbiterx AGENTS.md is an instructions file published in the GitHub repository NeelPrime/arbiterx (2 stars, last pushed 1mo ago), licensed Apache-2.0. It adds 345 tokens to every session, about $0.0017 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 2026-08-31.
Other instructions, from other repositories
procoder tests.instructions.md
Instructions for azrtydxb/procoder, covering tests in procoder, every test carries its mutation, assertions, portability and fixtures.
procoder copilot-instructions.md
Instructions for azrtydxb/procoder, covering procoder, the contract, the work chain, which command, right now and what you talk yourself into.
procoder go.instructions.md
Instructions for azrtydxb/procoder, covering go in procoder, no silent green, p-control: the binary prints, the agent writes, paths and arguments and exit codes are the public interface.
procoder workflows.instructions.md
Instructions for azrtydxb/procoder, covering ci in procoder, pins, actions and honesty in a job.
sdlc-quality copilot-instructions.md
Instructions for zircote-plugins/sdlc-quality, covering copilot instructions for sdlc plugin, project overview, repository structure, key conventions and rfc 2119 terminology.
sdlc-quality AGENTS.md
Instructions for zircote-plugins/sdlc-quality, covering agents.md - openai codex guidelines, project identity, quick context, agent guidelines and 1. understand the domain model.