cc-steer import

Import cc-factory decision outcomes as judged-pair candidates.

cc-steer import [OPTIONS] BATCH

Each decision lands as an ordinary feedback candidate keyed by a content digest of its source and external id, so re-importing the same batch is a no-op, and flows through the normal triage -> audit -> refine pipeline: the LLM gate re-judges every row rather than trusting the imported label. With --dry-run the batch is classified and reported without writing anything.

Full --help output

Usage: cc-steer import [OPTIONS] BATCH

  Import cc-factory decision outcomes as judged-pair candidates.

  Each decision lands as an ordinary feedback candidate keyed by a content
  digest of its source and external id, so re-importing the same batch is a
  no-op, and flows through the normal triage -> audit -> refine pipeline: the
  LLM gate re-judges every row rather than trusting the imported label. With
  --dry-run the batch is classified and reported without writing anything.

Options:
  --dry-run  Preview what would land, including dedup skips, without writing.
  --db FILE  Database path. Defaults to ~/.cc-steer/feedback.db.
  --help     Show this message and exit.

Arguments

BATCH: FILE: Required.

Options

--dry-run: Preview what would land, including dedup skips, without writing.
--db: FILE: Database path. Defaults to ~/.cc-steer/feedback.db.