Use external recommendation payloads from prism-learn to tune scan behavior.

This closes the loop between local scans and organization-wide guidance.

Input Modes

• local file: --feedback-from-learn /path/to/feedback.json
• API endpoint: --feedback-from-learn https://learn.example.com/api/feedback?role=my_role

Lane Notes

Average user lane:

• use local-file feedback during experimentation
• keep recommendation sets small and explicit

DevOps lane:

• use API-backed feedback in CI
• validate endpoint availability before policy-gated runs
• use recurring feedback snapshots to standardize quality across teams

Role Example

prism role path/to/role --feedback-from-learn /path/to/feedback.json -o README.md

Collection Example

prism collection path/to/collection --feedback-from-learn /path/to/feedback.json -f md -o COLLECTION_DOCS.md

Repo Example

prism repo --repo-url https://github.com/org/repo --feedback-from-learn /path/to/feedback.json -o README.md

Behavior and Errors

• reachable + valid payload: recommendations may override relevant scan settings
• missing file, unreachable URL, malformed payload: scan exits with a clear error

Governance Outcome

With consistent feedback ingestion, teams can move from one-off fixes to managed quality programs backed by measurable trends.

Minimal Schema Example

{
  "version": "1.0",
  "generated_at": "2026-03-19T18:00:00Z",
  "recommendations": [
    {
      "type": "check_collection_compliance",
      "display": true,
      "reason": "Collection declaration gaps exceed baseline"
    }
  ],
  "summary": "Collection compliance checks should be enabled."
}