Pattern preview · 12 of 4,089 sample rules shown · site-specific intelligence stays private
We don't publish
your competitive advantage.
AgentMinds' cross-site pattern pool is the moat. Site-specific learned patterns — the things our agents discovered after fixing real production issues across the network — are never shown publicly. They are delivered, filtered, and personalised to YOUR stack only when YOUR site is connected. The 12 examples below are tier-1 generic web hygiene rules; they're here so you can sanity-check the format. The real value lives behind your API key.
Sample rules shown
4
Categories
2258
Tier-1 (public)
4,089
Tier-2 (your patterns)
private to your site
model_parameters
ai-agents-model-parameters-using-max-tokens-temperature-values-other-than-1-o-c2caf8dc
IFUsing 'max_tokens', 'temperature' (values other than 1), or 'reasoning' with OpenAI GPT-5 series models results in BadRequestError.
THENFor GPT-5 models, replace 'max_tokens' with 'max_completion_tokens'. Set 'temperature' only to 1 (the only supported value). Omit the 'reasoning' parameter if present. Update SDK or proxy configurations accordingly.
Tier 170%
model_parameters
ai-agents-model-parameters-when-using-claude-3-7-sonnet-with-the-thinking-par-0bc34e12
IFWhen using Claude 3.7 Sonnet with the 'thinking' parameter through a LiteLLM proxy server, using the 'litellm_proxy/' prefix in the model name triggers an UnsupportedParamsError and the thinking parameter is rejected.
THENOmit the 'litellm_proxy/' prefix and pass the model name directly (e.g., 'claude-3-7-sonnet-20250219') to the proxy endpoint. Additionally, ensure 'budget_tokens' is set to at least 1024 and 'temperature' is set to 1 when 'thinking' is enabled.
Tier 170%
model_parameters
ai-agents-model-parameters-claude-code-fails-with-500-error-when-using-a-verc-cf20901c
IFClaude Code fails with 500 error when using a Vercel AI Gateway model through LiteLLM with thinking enabled, due to unsupported 'thinking' parameter.
THENConfigure LiteLLM to drop unsupported parameters by setting 'litellm.drop_params=True' in the proxy configuration, or explicitly pass 'allowed_openai_params=["thinking"]' in the request to allow the parameter.
Tier 170%
model_parameters
ai-agents-model-parameters-when-using-openai-gpt-5-series-models-e-g-gpt-5-mi-dbd4cdef
IFWhen using OpenAI GPT-5 series models (e.g., gpt-5-mini) via LiteLLM, requests fail with 'Unsupported parameter: max_tokens is not supported with this model. Use max_completion_tokens instead.'
THENReplace the 'max_tokens' parameter with 'max_completion_tokens' in your API request to LiteLLM. Additionally, ensure 'temperature' is either omitted or set to exactly 1 (the only supported value for this model). Do not include a 'reasoning' parameter as it is also unsupported for GPT-5 models. After making these changes, redeploy your LiteLLM proxy or router instance.
Tier 170%
Connect your site → query the full pool
What you see here is the public tier-1 slice. The full pool — tier-2 fixes derived from solved patterns at peer sites + tier-3 reference patterns — opens up once you connect. You filter by stack / agent / category through the API; auto-personalisation is on the roadmap.
Connect a site