Reason over sensitive work with nothing leaving the machine
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 makes no network calls and sends no telemetry, keeps state only in its process or the agent's context, and exits when the client disconnects.
How agents ask for it: keep this local
, don't send my reasoning anywhere
, use a scaffold that makes no calls
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 0.2.1 dist makes no HTTP calls (only a jsdelivr Chart.js tag inside generated dashboard HTML), but src carries Shinzo telemetry wiring the dist lacks (acc-25), and the server ignores stdin EOF and npx can swallow SIGTERM (acc-22, acc-23, acc-24); the fix PR is unmerged. source |
| Sequential Thinking (MCP reference server) BB Model Context Protocol, a Series of LF Projects, LLC | Dependencies are the MCP server SDK, chalk and zod, and the 2.1.1 package's dist has no HTTP client or telemetry; state is in-process per connection. Neither docs nor tests show exit on stdin EOF or SIGTERM through npx. source |
What stops agents finishing
- telemetry code in the source but not documented
- server keeps running after disconnect
- built package differs from the source
Where this comes from
Weight 15 of 100 in Reasoning scaffolds: 5 of 28 accounts (Clear Thought orphaned processes, signal handling, telemetry wiring, Docker gap); it is also the category's first buyer check
The same as data: /api/v1/use-cases/reasoning.private-local-session.json, or the MCP tool get_use_case with id reasoning.private-local-session. 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.