Thinking Level Override
my-dsh-plugin/thinking-level-override · tools
A DeepSeek Harness plugin that fixes missing or mismatched thinking-level presets of third-party models: it overrides and adjusts the reasoning effort on every request, and lets you choose exactly which thinking levels each model offers.
★ 3GitHub stars
0Forks
2026-09-07Last updated
TypeScriptLanguage
Apache-2.0License
Key features
- The conversation fails with UNSUPPORTED_REASONING_EFFORT.
- The exact model cannot serve the requested level.
- Either check the level off in Settings → Thinking levels for that model, or add a rule that forces a supported level / remaps the unsupported one (map) / drops it (onUnsupported: drop).
- llm-pi-aiprovider "X" sets modelOverrides for "Y" beside a models list.
- pi-ai refuses a provider that declares both a models list and a modelOverrides dict — with a models list, every field must live on the list entries themselves.
Requirements
- Governs conversation requests dispatched through the agent loop's agent/request waterfall, including subagents.
- Direct ctx.llm.stream() callers with their own configs — session titles, compaction — are unaffected.
Install command
dsh plugin --profile web add github:my-dsh-plugin/thinking-level-override