Skip to main content
Gets the learned ICP scoring model (which signals predict a win, their weight, lift, and the calibration gap) as a ready-to-write markdown block, so you can write it back into the user’s own ICP file. This is the payoff of the symbiosis: their file keeps the words, Nous keeps the model, and this writes the model under their words.
Direction is graph to file, the reverse of sync_icp. Call this after sync_icp or train_icp_model, then write the block back into the file.

​
Parameters

None. The tool reads the current model and the recorded write-back path.

​
Returns

The model as a block plus the file to write it into:
If there’s no model yet:

​
When to call it

  • After sync_icp → write the seeded model back under the user’s words
  • After train_icp_model → refresh the file with the model trained on real outcomes
When you write the block, replace whatever sits between the <!-- nous:icp start --> and <!-- nous:icp end --> markers, and never edit inside the markers by hand, because this tool regenerates them. Everything outside the markers is the user’s, so leave it untouched.

​
Backed by

GET /v2/workspace/icp/model. The agent setup surface. The server renders the block so the format stays controlled centrally.