Give a model or host without extended thinking a scratchpad
No company's docs show a full fit yet. The listings below do the job with a limit, or not at all.
What counts as done
The scaffold loads in that host within its tool-description and transport limits, the model writes several thoughts before answering, and nothing in the scaffold's output trips the provider's moderation.
How agents ask for it: make it reason before answering
, this model doesn't think, give it room
, force a thinking step
Does the job, with a limit 2
It does the job, but something in the docs or a user's account will get in an agent's way. The limit is named in each row.
| Listing | Why |
|---|---|
| Clear Thought 1.5 E Waldzell AI | The tool description is 642 characters, inside a 1,024 cap. The npm bin is stdio only; HTTP exists only as a Smithery build with no hosted URL documented, and per-operation fields sit in a guide resource a small model must read first. source |
| Sequential Thinking (MCP reference server) BB Model Context Protocol, a Series of LF Projects, LLC | Description trimmed to 1,009 characters in 2.1.0, under OpenAI's 1,024 cap (acc-16 fixed). Stdio only (npx or Docker), so ChatGPT-style remote-only hosts can't reach it (acc-09), and the GPT-5.5 moderation flag (acc-15) was closed not planned. source |
What stops agents finishing
- tool description longer than the provider allows
- host takes only remote MCP while the scaffold is stdio only
- provider moderation flags the scaffolded prompt
Where this comes from
Weight 20 of 100 in Reasoning scaffolds: 5 of 28 accounts (OpenAI description limit, GPT-5.5 moderation flag, ChatGPT transport, adaptive-thinking complaint, editor setup)
The same as data: /api/v1/use-cases/reasoning.scratchpad-without-native-thinking.json, or the MCP tool get_use_case with id reasoning.scratchpad-without-native-thinking. Other jobs in this category: /api/v1/niches?category=reasoning.
For companies
Do agents find, use and choose your tools?
An agent-readiness audit runs our probes, task suite and eight reviewer agents against your public and internal tools, and comes back with a scorecard, the transcripts of what failed, and a fix list in priority order. From $2,500, re-run included. We never take payment to move a rank. We do help companies earn one.