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/kiwunaka/codexrs/agents-mdgit clone --depth 1 https://github.com/Kiwunaka/codexRSWhat 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 | $0.00825 | $0.00825 |
| Opus 5 | $0.00413 | $0.00413 |
| Sonnet 5 | $0.00165 | $0.00165 |
| Haiku 4.5 | $0.00082 | $0.00082 |
Grade A, and why
codexRS 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.
How it starts
The opening of the file, as written. The whole thing — 68 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Project rules
- Build a native Rust replacement for Codex Desktop. Do not add Electron, Tauri, Wry, WebView, Node.js, or browser-runtime dependencies.
- Treat stable
26.721.3996.0as a behavioral reference, not a runtime dependency. - Never modify the installed MSIX. Never open the live
CODEX_HOMESQLite, JSONL, auth, or log files directly. Runtime access to the default~/.codexgoes through a supervised officialcodex app-serverprocess; development and tests use an isolatedCODEX_HOME. - Bound every external frame, event, log, queue, and history query. No
unbounded
read_to_string, JSONL line reads, or startup scans. - Keep platform process management behind
codex-platform. On Windows, prefer Job Objects and graceful cancellation; never implement pollingtaskkillloops. - Keep codexRS-owned storage single-writer and paginated. Existing Codex data
remains app-server-owned and is queried through bounded, paginated protocol
methods (
thread/listmust setuseStateDbOnly). Use snapshots for direct import, recovery, and fixtures; never share direct file access. - Run
cargo fmt --all --check,cargo clippy --workspace --all-targets, andcargo test --workspacebefore handing off changes.
Proportional Engineering
- Apply KISS, YAGNI, and the Pareto principle. Make the smallest maintainable change that satisfies the explicit acceptance criteria and current evidence; prefer existing patterns and code paths.
- Do not add speculative abstractions, dependencies, compatibility layers, fallbacks, configuration, cleanup, documentation, or future-proofing outside the assigned scope.
- Keep verification proportional. Add or update only the smallest focused tests needed to prove changed behavior or prevent a concrete observed regression. Do not add redundant unit/integration/E2E coverage, exhaustive edge-case matrices, broad regression suites, or unrelated test refactors unless the task, affected shared contract, or observed failure requires them.
- Keep security work proportional to the actual trust boundary and concrete threat model. Preserve mandatory safeguards and fix vulnerabilities introduced or exposed by the task, but do not add speculative hardening, new security frameworks, or unrelated defenses without evidence or an explicit requirement.
- Before expanding scope, identify the concrete acceptance criterion, failure, or risk that requires it. If none exists, omit the extra work. If expansion would materially change the solution, request owner direction first.
- These proportionality rules do not authorize skipping checks explicitly required by the selected router row, current repository contracts, or release gates applicable to the changed behavior.
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 · 68 lines · 825 tokens per session scan A 29413fa6dcbb
codexRS AGENTS.md is an instructions file published in the GitHub repository Kiwunaka/codexRS (5 stars, last pushed 18d ago), licensed Apache-2.0. It adds 825 tokens to every session, about $0.0041 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
codex-trace AGENTS.md
Instructions for PixelPaw-Labs/codex-trace, covering claude.md, commands, dev, lint and format.
codex-usage-desktop AGENTS.md
Instructions for itvincent-git/codex-usage-desktop, covering agents.md, 1. think before coding, 2. simplicity first, 3. surgical changes and 4. goal-driven execution.
codex-usage-widget AGENTS.md
Instructions for ognjeeen/codex-usage-widget, covering agents.md, project, scope and decision discipline, authorization and repository safety and architecture boundaries.
agent-usage-desktop AGENTS.md
Instructions for hongshuo-wang/agent-usage-desktop, covering agents.md, build & run, desktop app (tauri), macos arm64 example and macos x8664.
Osverse AGENTS.md
AGENTS.md instructions for Oswald-Hao/Osverse, covering osverse repository rules for coding agents and protected-branch workflow.
CodexCliPlus AGENTS.md
Instructions for C4AL/CodexCliPlus, covering agents.md and repository rules.