mirror of
https://github.com/Nutlope/hallmark.git
synced 2026-08-14 12:35:33 +02:00
Arms: Anthropic baseline, GLM on Together, Kimi on OpenRouter (DeepSeek and Qwen staged, disabled). Six briefs incl. metric-temptation, component-scope, and vision-probe probes. Floor-only vs full skill packs to measure prompt bloat. run-matrix orchestrates gen, shots, score, judge, report; the repair loop feeds sloplint FAILs back to the model.
50 lines
1.5 KiB
JSON
50 lines
1.5 KiB
JSON
{
|
|
"arms": [
|
|
{
|
|
"id": "fable-baseline",
|
|
"kind": "anthropic",
|
|
"model": "claude-fable-5",
|
|
"env": "ANTHROPIC_API_KEY"
|
|
},
|
|
{
|
|
"id": "glm52-together",
|
|
"kind": "openai-compat",
|
|
"base": "https://api.together.xyz/v1",
|
|
"model": "zai-org/GLM-5.2",
|
|
"env": "TOGETHER_API_KEY",
|
|
"temperature": 0.6,
|
|
"extra": { "chat_template_kwargs": { "thinking": false } }
|
|
},
|
|
{
|
|
"id": "k3-openrouter",
|
|
"kind": "openai-compat",
|
|
"base": "https://openrouter.ai/api/v1",
|
|
"model": "moonshotai/kimi-k3",
|
|
"env": "OPENROUTER_API_KEY",
|
|
"omitTemperature": true,
|
|
"extra": { "reasoning": { "effort": "medium" } }
|
|
},
|
|
{
|
|
"id": "deepseek-openrouter",
|
|
"kind": "openai-compat",
|
|
"base": "https://openrouter.ai/api/v1",
|
|
"model": "deepseek/deepseek-v3.2",
|
|
"env": "OPENROUTER_API_KEY",
|
|
"enabled": false
|
|
},
|
|
{
|
|
"id": "qwen-openrouter",
|
|
"kind": "openai-compat",
|
|
"base": "https://openrouter.ai/api/v1",
|
|
"model": "qwen/qwen3-coder-next",
|
|
"env": "OPENROUTER_API_KEY",
|
|
"enabled": false
|
|
}
|
|
],
|
|
"keyFiles": [
|
|
"/Users/youssef/Desktop/.env.local",
|
|
"/Users/youssef/Inspo-design/.env"
|
|
],
|
|
"keyFileNotes": "Verified 2026-07-23: Desktop/.env.local defines ANTHROPIC_API_KEY and TOGETHER_API_KEY; Inspo-design/.env defines TOGETHER_API_KEY. OPENROUTER_API_KEY is NOT present in either file yet; the OpenRouter arms need it exported in the shell or added to a key file before they can run."
|
|
}
|