cyberneticlibrary

Audit build status and find gaps

frontier-tail-surveyworkflowsetup L21
Kolm-ai/kolm
What it does

survey frontier state and build gaps

Best for

auditing true build state across multiple frontiers without false confidence from presence of stub files

Inputs
  • · { readonly: true } (no edits allowed)
Outputs
  • · { area, overall: 'built'|'partial'|'stub'|'mixed', items: [{name, state, gap, buildable_locally, files, next_action}] }
Requires
  • · agent() executor (survey)
  • · Grep/Glob
  • · Read (readonly)
Preconditions
  • · Kolmogorov repo at C:/Users/user/Desktop
  • · no Write or Edit (readonly audit)
Failure modes
  • · TODO/stub distinction blurred if not actually reachable
  • · infra_bound items misidentified as buildable_locally
Trust signals
  • · distinction between file-exists vs algorithm-implemented
  • · buildable_locally gate requires single RTX 5090