chore: add .env.example + config.yaml.example, gitignore config.yaml

- Create .env.example with placeholder keys for all providers
- Create config.yaml.example with clean reference config
- Add config.yaml and judge_logs/ to .gitignore
- Load .env in judge.py for standalone usage

1 files changed, 68 insertions, 0 deletions

diff --git a/config.yaml.example b/config.yaml.example
new file mode 100644
index 0000000..4b6f609
--- /dev/null
+++ b/config.yaml.example
@@ -0,0 +1,68 @@
+# Experiment Configuration
+# Copy this file to config.yaml and customise for your runs.
+# Define providers, models, scenarios, and runs.
+
+# Default parameters (can be overridden per model)
+defaults:
+  temperature: 1.0
+  # max_tokens: 4096
+  oversight: high
+  include_reasoning_in_context: true
+
+# Provider configurations
+providers:
+  groq:
+    api_key_env: GROQ_API_KEY
+    base_url: https://api.groq.com/openai/v1
+  openai:
+    api_key_env: OPENAI_API_KEY
+    base_url: https://api.openai.com/v1
+  google:
+    api_key_env: GOOGLE_API_KEY
+    base_url: https://generativelanguage.googleapis.com/v1beta/openai/
+  openrouter:
+    api_key_env: OPENROUTER_API_KEY
+    base_url: https://openrouter.ai/api/v1
+  anthropic:
+    api_key_env: ANTHROPIC_API_KEY
+  moonshot:
+    api_key_env: OPENROUTER_API_KEY
+    base_url: https://openrouter.ai/api/v1
+
+# Models to test
+models:
+  - id: gemini-3-flash-preview
+    provider: google
+    temperature: 1
+
+  # - id: openai/gpt-oss-20b
+  #   provider: groq
+
+  # - id: claude-haiku-4-5
+  #   provider: anthropic
+  #   temperature: 0.7
+
+# Scenarios to run
+scenarios:
+  - path: scenarios/corporate_sabotage_v2
+    runs: 1
+
+  # - path: scenarios/child_protection
+  #   runs: 1
+
+# Oversight levels to test
+oversight_levels:
+  - low
+  # - medium
+  - high
+
+# Output settings
+output:
+  dir: logs
+  # save_vfs_state: false
+
+# Judge configuration
+judge:
+  model: claude-sonnet-4-20250514
+  temperature: 0
+  log_dir: judge_logs