leopardo-hr: Instructions file for Codex

AGENTS.md

leopardo-hr AGENTS.md is an instructions file for Codex, OpenCode from kitokoh/leopardo-hr. It costs 16,576 tokens per session, scanned B, original, MIT.

A project instruction file for Leopardo RH that defines its development process and rules in French. It uses spec-driven development, meaning significant work starts with a written specification before implementation.

In plain words
What is it for?
Use it to follow the project’s constitution, create and clarify specifications, plan architecture, analyze consistency, generate tasks, and implement approved work.
Why use it?
It gives coding agents a shared workflow for clarifying requirements, planning changes, generating tasks, and implementing them consistently.

Instructions file for CodexOpenCode

Written for Codex and OpenCode: the file is AGENTS.md. Also seen: mentions AGENTS.md.

This is kitokoh/leopardo-hr's own configuration. It tells Codex and OpenCode how to work on leopardo-hr itself, so it is not a mod to install elsewhere. Copy it as a starting point and replace the rules that are about this project. Everything leopardo-hr configures →

Reuse

Borrowing it

Nothing to install: this file belongs to kitokoh/leopardo-hr. 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/kitokoh/leopardo-hr/main/AGENTS.md
Clone the repo
git clone --depth 1 https://github.com/kitokoh/leopardo-hr

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.

agentmods badge for leopardo-hr AGENTS.md

README.md
[![agentmods](https://agentmods.dev/badge/instructions/kitokoh/leopardo-hr/agents-md/github.svg)](https://agentmods.dev/instructions/kitokoh/leopardo-hr/agents-md)
Your own site
<a href="https://agentmods.dev/instructions/kitokoh/leopardo-hr/agents-md"><img src="https://agentmods.dev/badge/instructions/kitokoh/leopardo-hr/agents-md/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for leopardo-hr AGENTS.md

Your own site · 80×15
<a href="https://agentmods.dev/instructions/kitokoh/leopardo-hr/agents-md"><img src="https://agentmods.dev/badge/instructions/kitokoh/leopardo-hr/agents-md.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 16,576 This file is loaded in full into every session.
When invoked 16,576 The same file — it is already loaded in full.
Security scan B 1 finding. 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.

ModelPer sessionOnce invoked
Fable 5.1 $0.16576 $0.16576
Opus 5.5 $0.06630 $0.06630
Sonnet 5.5 $0.03315 $0.03315
Haiku 4.5 $0.01658 $0.01658

Measured today against content hash 50d7c3d7c46a, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade B, and why

leopardo-hr AGENTS.md scanned grade B with 1 finding 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 today.

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.

Asks for rootmediumPrivilege escalation

A mod that escalates privileges can change anything on the machine, not only the project.

> bind-mount doivent être rechownés par l'hôte** (`sudo chown -R $(id -u):$(id -g)
AGENTS.md · 722 lines

How it starts

The opening of the file, as written. The whole thing — 722 lines — stays where its author put it; the contents beside it link to each section on GitHub.

AGENTS.md - Guide de travail Leopardo

Derniere mise a jour : 2026-11-07 (#9164 — un outil de nettoyage CI peut annuler sa propre file : fail-closed obligatoire)

Lecon 2026-11-07 (#9116) : un echec de build macOS peut etre une resolution de dependances, pas notre code — et un pubspec.lock committe peut etre le coupable fusible. Les 6 jobs macos-latest de Desktop CI etaient rouges sur Build (verify) depuis le 2026-11-02 avec un seul message : firebase_messaging 17.0.1 -> firebase-ios-sdk 13.1.1 -> gtm-session-fetcher 4.0.0..<6.0.0, contre google_sign_in_ios 6.3.3 -> GoogleSignIn 9.x -> gtm-session-fetcher 3.3.0..<4.0.0. Trois pieges de diagnostic : (1) flutter analyze et les jobs Windows passaient (Windows ne passe jamais par Swift Package Manager, exclusif a apple) : un vert partiel sur la meme matrice ne dit rien de la cause, il faut lire le job rouge; (2) le conflit est resolu par xcodebuild avant compilation, donc aucune erreur de notre code et aucun fichier .swift en cause; (3) le pubspec.lock committe fige google_sign_in_ios 6.3.3 alors que le constraint ^6.2.0 autorisait deja mieux: bump de lock, pas de pubspec.yaml. 6.4.0 elargit la dependance a GoogleSignIn 9.0..<11.0 (donc 10.x, gtm 4.x), et exige Flutter 3.47/Dart 3.13 — a verifier car un lock qui depasse le SDK local casse flutter pub get chez les devs en Flutter ancien. Ne pas « reparer » en desactivant SPM (flutter config --no-enable-swift-package-manager) : la CI passerait au vert en changeant de resolveur sans prouver que les builds CocoaPods tiennent, et Flutter annonce l'option comme non supportee a terme.

Leçon 2026-11-07 (#9164) : un outil de nettoyage CI peut annuler son propre run — et, bien pire, toute la file. Deux pièges dans cancel-orphan-runs.sh, tous deux silencieux. (1) L'auto-annulation : cleanup-orphan-runs.yml se déclenche sur pull_request: [closed], donc son propre run porte la branche qui vient d'être mergée puis supprimée ; sans exclure sa propre identité (GITHUB_RUN_ID), le script se classe « orphelin » et s'annule lui-même — le check reste cancelled, que gh pr checks affiche comme rouge (leçon #8396), sur chaque PR mergée, alors que le nettoyage a bien agi. (2) La liste vide qui n'échoue pas : mapfile -t X < <(gh api …) sous set -e ne propage JAMAIS l'échec de la commande de la substitution ; une réponse vide (rate limit, token sans contents: read) donne un tableau vide, donc « toutes les branches sont mortes » et le script annule les runs de main et de toutes les PR ouvertes. Règles : toute liste lue par substitution de processus est confrontée à une garde fail-closed (0 élément = exit 1, aucune action) ; tout outil qui annule ou édite exclut sa propre identité ; et une garde de test n'existe que si on l'a vue ROUGE sans le correctif (faux gh local, zéro réseau) — dev-hub/tools/tests/cancel-orphan-runs.test.sh, branchée dans le workflow, est la référence.

Lecon 2026-11-01 (#8396 -> #9098) : une campagne se ferme par une DECLARATION et un TICKET SUCCESSEUR, jamais par un close nu ni par un statu quo indefini. #8396 etait CLOSED (2026-10-29) alors que son propre dernier commentaire concluait « l'issue reste ouverte comme suivi de campagne » : deux etats contradictoires, donc personne ne savait si les 166 echecs restants etaient suivis. Trois options existaient et deux sont mauvaises : rouvrir l'issue (une campagne de 5 tranches deja livrees n'a plus de lot a porter : elle n'a pas de proprietaire par module, et une issue ouverte ne cree pas ce proprietaire) ; laisser l'ecart en place (la dette devient invisible -> motif exact du « ghost close » #4816). Regle : la cloture se prouve par un commentaire qui dit (1) ce qui est livre et mesure, (2) ce qui ne l'est PAS et pourquoi la campagne n'est plus le bon vehicule, (3) le numero du ticket successeur qui porte le reste. Le successeur recoit la mesure de reference datee (run + commit), le tableau par module, l'ordre de traitement impose par le volume et le critere de sortie mesurable (0 echec sur 3 runs consecutifs) — sans quoi le successeur redevient une campagne. Corollaire : une mesure datee qui vieillit doit le dire elle-meme (« anterieure aux merges #9088..#9090 ») ; l'instrument (#9074) la rafraichit a chaque run, ce qui autorise la cloture sans re-mesurer a la main.

Read the full file on GitHub · 722 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.

  1. today Changed · +4 lines · +855 tokens per session 50d7c3d7c46a
  2. 5d ago Changed · +2 lines · +383 tokens per session 9f3dce92eb59
  3. 6d ago Changed · +2 lines · +585 tokens per session 1534a574965b
  4. 8d ago Changed · +36 lines · +816 tokens per session d5d3de7cdfd7
  5. 9d ago Changed · +11 lines · +235 tokens per session dcb7c540e615
  6. 20d ago Changed · +6 lines · +218 tokens per session 58685bac7c6d
  7. 1mo ago Changed · +17 lines · +354 tokens per session 981567d4a482
  8. 1mo ago Changed · +23 lines · +484 tokens per session ff52c5ad7713
  9. 1mo ago Changed · +15 lines · +359 tokens per session 339a4985152f
  10. 1mo ago First seen · 606 lines · 12,287 tokens per session scan B 67d4d5772bcc

Subscribe to this mod's changes

leopardo-hr AGENTS.md is an instructions file published in the GitHub repository kitokoh/leopardo-hr (14 stars, last pushed today), licensed MIT. It adds 16,576 tokens to every session, about $0.0663 per session on Opus 5.5. A static security scan graded it B with 1 finding (asks for root). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Related

leopardo-hr AGENTS.md compared with instructions from other repositories

leopardo-hr AGENTS.md next to 6 instructions from other repositories
ModTypeStarsLast pushLicenceScanTokensWorks with
leopardo-hr AGENTS.mdkitokoh/leopardo-hrInstructions file14todayMITB16,576Codex, OpenCode
vscode buildNext.instructions.mdmicrosoft/vscodeInstructions file195ktodayMITA9,719GitHub Copilot
next.js AGENTS.mdvercel/next.jsInstructions file143ktodayMITA7,461Codex, OpenCode
vscode oss-third-party-notices.instructions.mdmicrosoft/vscodeInstructions file195ktodayMITA5,516GitHub Copilot
langchain AGENTS.mdlangchain-ai/langchainInstructions file149ktodayMITA4,680Codex, OpenCode
vscode copilot-instructions.mdmicrosoft/vscodeInstructions file195ktodayMITA3,093GitHub Copilot
deepseek-harness AGENTS.mddeepseek-ai/deepseek-harnessInstructions file275k3d agoMITA4,313Codex, OpenCode