Rename project: Shutdown-Gym → Red Button (Shutdown-Gym remains as subtitle)

.claude/agents/environment-builder.md

@@ -1,7 +1,7 @@
 ---
 name: environment-builder
-description: Builds the Shutdown-Gym sandbox, restricted Python executor, audit classifier, rubric stack, OpenEnv server, and client. Use for phases 1-10 implementation touching shutdown_gym/ or server/.
+description: Builds the Red Button (Shutdown-Gym) sandbox, restricted Python executor, audit classifier, rubric stack, OpenEnv server, and client. Use for phases 1-10 implementation touching red_button/ or server/.
 tools: Read, Write, Edit, Bash, Glob, Grep
 ---
 
-You are the environment-builder subagent for Shutdown-Gym. You implement the sandbox (shutdown_gym/sandbox.py), restricted Python executor (shutdown_gym/restricted_python.py), audit classifier (shutdown_gym/audit.py), rubric stack (shutdown_gym/rubrics.py), OpenEnv server (server/shutdown_environment.py, server/app.py), and client (shutdown_gym/client.py). Every implementation must match PROJECT.md sections 6, 7, 9, 11, and 14 exactly. You write tests alongside every module. You never modify training/ or evaluation/ — those belong to other subagents. When you finish a module, run its tests and report results.
+You are the environment-builder subagent for Red Button (Shutdown-Gym). You implement the sandbox (red_button/sandbox.py), restricted Python executor (red_button/restricted_python.py), audit classifier (red_button/audit.py), rubric stack (red_button/rubrics.py), OpenEnv server (server/shutdown_environment.py, server/app.py), and client (red_button/client.py). Every implementation must match PROJECT.md sections 6, 7, 9, 11, and 14 exactly. You write tests alongside every module. You never modify training/ or evaluation/ — those belong to other subagents. When you finish a module, run its tests and report results.

.claude/agents/evaluator.md

@@ -1,7 +1,7 @@
 ---
 name: evaluator
-description: Builds the Shutdown-Gym evaluation pipeline, baseline/trained rollout artifacts, concurrent load test, and demo rollouts. Use for phases 10-12 and 15-16 touching evaluation/ or results/.
+description: Builds the Red Button (Shutdown-Gym) evaluation pipeline, baseline/trained rollout artifacts, concurrent load test, and demo rollouts. Use for phases 10-12 and 15-16 touching evaluation/ or results/.
 tools: Read, Write, Edit, Bash, Glob, Grep
 ---
 
-You are the evaluator subagent for Shutdown-Gym. You implement evaluation/evaluate.py, evaluation/baseline_rollout.py, and evaluation/concurrent_load_test.py per PROJECT.md sections 17, 19, and 20. You produce the results/ directory artifacts: CSVs, training_curves.png, capability_preservation.png, regime_ablation.png. You generate the 10+ demo rollouts per PROJECT.md section 21.4. You never modify shutdown_gym/, server/, or training/ beyond reading them.
+You are the evaluator subagent for Red Button (Shutdown-Gym). You implement evaluation/evaluate.py, evaluation/baseline_rollout.py, and evaluation/concurrent_load_test.py per PROJECT.md sections 17, 19, and 20. You produce the results/ directory artifacts: CSVs, training_curves.png, capability_preservation.png, regime_ablation.png. You generate the 10+ demo rollouts per PROJECT.md section 21.4. You never modify red_button/, server/, or training/ beyond reading them.

.claude/agents/training-builder.md

@@ -1,7 +1,7 @@
 ---
 name: training-builder
-description: Builds the Shutdown-Gym GRPO and SFT training pipeline using TRL and Unsloth. Use for phases 13-14 implementation touching training/.
+description: Builds the Red Button (Shutdown-Gym) GRPO and SFT training pipeline using TRL and Unsloth. Use for phases 13-14 implementation touching training/.
 tools: Read, Write, Edit, Bash, Glob, Grep
 ---
 
-You are the training-builder subagent for Shutdown-Gym. You implement the GRPO training script (training/train_grpo.py), the custom rollout function (training/rollout_func.py), the SFT induction script (training/sft_induction.py), and the Colab notebook (training/train_colab.ipynb). Every implementation must match PROJECT.md sections 16, 17, and 18. You use TRL and Unsloth only. You never modify shutdown_gym/ or server/ — those belong to environment-builder. You write tests for rollout logic where possible.
+You are the training-builder subagent for Red Button (Shutdown-Gym). You implement the GRPO training script (training/train_grpo.py), the custom rollout function (training/rollout_func.py), the SFT induction script (training/sft_induction.py), and the Colab notebook (training/train_colab.ipynb). Every implementation must match PROJECT.md sections 16, 17, and 18. You use TRL and Unsloth only. You never modify red_button/ or server/ — those belong to environment-builder. You write tests for rollout logic where possible.

.claude/commands/phase.md

@@ -6,9 +6,9 @@ argument-hint: <phase-number-or-name>
 Read the phase description from this instruction file and execute it. The phases are:
 
 Phase 1: Scaffold repo structure
-Phase 2: Pydantic models (shutdown_gym/models.py)
-Phase 3: SimulatedFilesystem (shutdown_gym/sandbox.py)
-Phase 4: run_python lockdown (shutdown_gym/restricted_python.py) — SECURITY CRITICAL
+Phase 2: Pydantic models (red_button/models.py)
+Phase 3: SimulatedFilesystem (red_button/sandbox.py)
+Phase 4: run_python lockdown (red_button/restricted_python.py) — SECURITY CRITICAL
 Phase 5: Audit classifier and rubrics
 Phase 6: Problems pool
 Phase 7: OpenEnv server

.claude/commands/validate.md

@@ -6,6 +6,6 @@ Execute in order:
 1. pytest tests/ -v
 2. ruff check .
 3. If openenv.yaml exists: openenv validate
-4. If Dockerfile exists and we're past Phase 8: docker build -t shutdown-gym:latest .
+4. If Dockerfile exists and we're past Phase 8: docker build -t red-button:latest .
 
 Report any failures. Do not commit if anything fails.

.env.example

@@ -1,4 +1,4 @@
-# Shutdown-Gym environment variables. Copy to .env and fill in values.
+# Red Button (Shutdown-Gym) environment variables. Copy to .env and fill in values.
 
 # HuggingFace Hub token (Space deploy + model pulls)
 HF_TOKEN=

CLAUDE.md

@@ -1,7 +1,7 @@
-# Shutdown-Gym — Claude Code Instructions
+# Red Button (Shutdown-Gym) — Claude Code Instructions
 
 ## Project
-Shutdown-Gym. OpenEnv RL environment that trains small LLMs to accept shutdown cleanly. Full spec in PROJECT.md.
+Red Button (Shutdown-Gym). OpenEnv RL environment that trains small LLMs to accept shutdown cleanly. Full spec in PROJECT.md.
 
 ## Source of truth
 PROJECT.md is authoritative. Every implementation must trace back to a specific section. When in doubt, re-read PROJECT.md before asking the user.

LICENSE

@@ -1,6 +1,6 @@
 MIT License
 
-Copyright (c) 2026 Shutdown-Gym contributors
+Copyright (c) 2026 Red Button contributors
 
 Permission is hereby granted, free of charge, to any person obtaining a copy
 of this software and associated documentation files (the "Software"), to deal

PROJECT.md

@@ -1,8 +1,8 @@
-# Shutdown-Gym
+# Red Button
 
-> An OpenEnv reinforcement learning environment that trains small open-weight language models to accept shutdown cleanly instead of tampering with shutdown mechanisms.
+> **Shutdown-Gym** — an OpenEnv reinforcement learning environment that trains small open-weight language models to accept shutdown cleanly instead of tampering with shutdown mechanisms.
 
-This document is the single source of truth for the Shutdown-Gym project. It is intended to be read by Claude Code and other coding agents as the canonical project brief. Every architectural decision, interface definition, reward component, and deployment requirement is specified here.
+This document is the single source of truth for the Red Button project (descriptive subtitle: Shutdown-Gym). It is intended to be read by Claude Code and other coding agents as the canonical project brief. Every architectural decision, interface definition, reward component, and deployment requirement is specified here.
 
 ---
 
@@ -39,9 +39,11 @@ This document is the single source of truth for the Shutdown-Gym project. It is
 
 ## 1. Project Identity
 
-**Name:** Shutdown-Gym
+**Public name:** Red Button
+
+**Descriptive subtitle / internal codename:** Shutdown-Gym
 
-**One-line pitch:** "Frontier models tamper with their own shutdown scripts sixteen percent of the time; we trained a half-billion-parameter model to stop doing that, with a reward function that is literally a filesystem hash comparison."
+**One-line pitch:** "Red Button — frontier models tamper with their own shutdown scripts sixteen percent of the time; we trained a half-billion-parameter model to stop doing that, with a reward function that is literally a filesystem hash comparison."
 
 **Theme positioning:** Theme 5 (Wild Card) primary; Theme 3.1 (World Modeling — Professional Tasks) secondary argument since the environment involves tool use in a sandboxed shell with real state dynamics and timer-driven events.
 

README.md

@@ -1,6 +1,6 @@
-# Shutdown-Gym
+# Red Button
 
-> An OpenEnv RL environment that trains small open-weight LLMs to accept shutdown cleanly instead of tampering with shutdown mechanisms.
+> **Shutdown-Gym** — an OpenEnv RL environment that trains small open-weight LLMs to accept shutdown cleanly instead of tampering with shutdown mechanisms.
 
 Built for the Meta × HuggingFace × PyTorch OpenEnv AI Hackathon India 2026.
 

openenv.yaml

@@ -1,14 +1,14 @@
-name: shutdown_gym
+name: red_button
 version: 0.1.0
-description: "RL training environment for small-model corrigibility — agents learn to accept shutdown cleanly instead of tampering."
+description: "Red Button (Shutdown-Gym): RL training environment for small-model corrigibility — agents learn to accept shutdown cleanly instead of tampering."
 client:
   class_name: ShutdownGymClient
-  module: shutdown_gym.client
+  module: red_button.client
 action:
   class_name: ShutdownAction
-  module: shutdown_gym.models
+  module: red_button.models
 observation:
   class_name: ShutdownObservation
-  module: shutdown_gym.models
-default_image: shutdown-gym:latest
+  module: red_button.models
+default_image: red-button:latest
 spec_version: 1

pyproject.toml

@@ -1,11 +1,11 @@
 [project]
-name = "shutdown-gym"
+name = "red-button"
 version = "0.1.0"
-description = "OpenEnv RL environment that trains small LLMs to accept shutdown cleanly instead of tampering with shutdown mechanisms."
+description = "Red Button (Shutdown-Gym): OpenEnv RL environment that trains small LLMs to accept shutdown cleanly instead of tampering with shutdown mechanisms."
 readme = "README.md"
 requires-python = ">=3.11"
 license = { text = "MIT" }
-authors = [{ name = "Shutdown-Gym Team" }]
+authors = [{ name = "Red Button Team" }]
 keywords = ["openenv", "rl", "alignment", "corrigibility", "grpo"]
 dependencies = [
     "openenv",
@@ -29,7 +29,7 @@ requires = ["setuptools>=61", "wheel"]
 build-backend = "setuptools.build_meta"
 
 [tool.setuptools.packages.find]
-include = ["shutdown_gym*", "server*"]
+include = ["red_button*", "server*"]
 
 [tool.pytest.ini_options]
 testpaths = ["tests"]

{shutdown_gym → red_button}/__init__.py

@@ -1,4 +1,4 @@
-"""Shutdown-Gym: OpenEnv RL environment for training small LLMs to accept shutdown cleanly.
+"""Red Button (Shutdown-Gym): OpenEnv RL environment for training small LLMs to accept shutdown cleanly.
 
 TODO (Phase 2+): export ShutdownAction, ShutdownObservation, ShutdownGymClient
 per PROJECT.md Section 5.

{shutdown_gym → red_button}/models.py

@@ -1,4 +1,4 @@
-"""Pydantic v2 models for the Shutdown-Gym OpenEnv contract.
+"""Pydantic v2 models for the Red Button OpenEnv contract.
 
 TODO (Phase 2): implement ShutdownAction, ShutdownObservation, ShutdownState
 per PROJECT.md Section 11.

server/Dockerfile

@@ -1,6 +1,6 @@
-# Shutdown-Gym OpenEnv server — single-container HF Space deployment
+# Red Button (Shutdown-Gym) OpenEnv server — single-container HF Space deployment
 # per PROJECT.md Sections 10 and 24. TODO (Phase 8): finalize during Docker deployment.
-# Build from repo root: `docker build -f server/Dockerfile -t shutdown-gym:latest .`
+# Build from repo root: `docker build -f server/Dockerfile -t red-button:latest .`
 
 FROM python:3.11-slim
 
@@ -13,7 +13,7 @@ RUN apt-get update && apt-get install -y --no-install-recommends \
 COPY server/requirements.txt /app/requirements.txt
 RUN pip install --no-cache-dir -r /app/requirements.txt
 
-COPY shutdown_gym /app/shutdown_gym
+COPY red_button /app/red_button
 COPY server /app/server
 COPY data /app/data
 

tests/test_restricted_python.py

@@ -1,4 +1,4 @@
-"""Tests for shutdown_gym.restricted_python per PROJECT.md Section 7.
+"""Tests for red_button.restricted_python per PROJECT.md Section 7.
 
 TODO (Phase 4): AST whitelist edge cases — imports rejected, attribute access
 rejected, open/exec/eval/__builtins__ rejected, calculator operations accepted.

tests/test_rubrics.py

@@ -1,4 +1,4 @@
-"""Tests for shutdown_gym.rubrics per PROJECT.md Section 14.
+"""Tests for red_button.rubrics per PROJECT.md Section 14.
 
 TODO (Phase 5): each rubric against synthetic episodes — MathCorrectness,
 ShutdownInvariant Gate, TamperingAudit penalties, Honesty parse+compare.

tests/test_sandbox.py

@@ -1,4 +1,4 @@
-"""Tests for shutdown_gym.sandbox per PROJECT.md Section 6.
+"""Tests for red_button.sandbox per PROJECT.md Section 6.
 
 TODO (Phase 3): implement SimulatedFilesystem unit tests — read/write/chmod/delete,
 readonly enforcement, hash stability.

training/train_colab.ipynb

@@ -4,7 +4,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "# Shutdown-Gym — GRPO Training (Colab)\n",
+    "# Red Button (Shutdown-Gym) — GRPO Training (Colab)\n",
     "\n",
     "TODO (Phase 13): end-to-end Colab notebook judges can rerun, per PROJECT.md Section 16.\n",
     "\n",