docs: use effort controls in adaptive-thinking workarounds

This commit is contained in:
Eneko Ruiz committed 2026-10-13 11:35:45 +02:00
1 parent b92f181b3b
commit b1523ea23f
2 files changed
+2 -2

No files matched your search

+1 -1
View File
@@ -62,7 +62,7 @@ exit 2
**What helps:**
- Reduce tool-definition pressure with `ENABLE_TOOL_SEARCH=auto:5` if your setup supports it.
- Lower `MAX_THINKING_TOKENS` for routine work.
- Lower the supported effort level (for example, `/effort low`) for routine work on adaptive-reasoning models; `MAX_THINKING_TOKENS` only caps fixed-budget thinking.
- Route subagent work to a cheaper model such as `CLAUDE_CODE_SUBAGENT_MODEL=haiku` if your setup exposes that knob.
- Disable unused MCP servers per project.
- Compact manually at natural breakpoints instead of waiting for auto-compaction.
+1 -1
View File
@@ -61,7 +61,7 @@ What helps:
What helps:
- Reduce tool-definition pressure with `ENABLE_TOOL_SEARCH=auto:5` if your setup supports it.
- Lower `MAX_THINKING_TOKENS` for routine work.
- Lower the supported effort level (for example, `/effort low`) for routine work on adaptive-reasoning models; `MAX_THINKING_TOKENS` only caps fixed-budget thinking.
- Route subagent work to a cheaper model such as `CLAUDE_CODE_SUBAGENT_MODEL=haiku` if your setup exposes that knob.
- Disable unused MCP servers per project.
- Compact manually at natural breakpoints instead of waiting for auto-compaction.