Governance & Learning
sate review-queue
Shows the REVIEW-class learning candidates a human needs to approve, reject, or defer.
What it does
Reads a candidates JSON (typically from emit-learning-candidates), classifies each via the promotion engine (AUTO / REVIEW / FORBIDDEN), and surfaces the REVIEW queue with one entry per pending candidate. Previously-rejected candidates are filtered out (the system never re-asks). Each entry shows the candidate, its content-addressed id, prior decisions, and a precedent weight (count of prior APPROVEs for the same candidate kind).
Who runs this
- →Learning-stewardship reviewers (the human approval step before a pattern enters the KB)
How to invoke
sate review-queue --run-id r1 --candidates candidates.json --project-root /path/to/laravel-appFlags
--run-id <id>Required. Deterministic run identifier.
--candidates <path>Required. Path to candidates JSON.
--project-root <path>Anchors --review-ledger default at <root>/review/reviewLedger.ndjson.
--review-ledger <path>Override ledger location.
What it produces
- ·JSON queue + summary on stdout
What it does NOT do
- ✖Does NOT apply any decision — purely a viewer
- ✖Does NOT show AUTO or FORBIDDEN classes (handled elsewhere)
Status
Engine tests
22
Status
Shipped
Category
Governance & Learning
Related pipelines
sate submit-decisionAppend a human decision (APPROVE / REJECT / DEFER) for a candidate to the append-only review ledger.
sate emit-learning-candidatesOpt-in: extracts STABLE_TESTHASH_PASS candidates from an execution trace for downstream review.
sate learningMulti-repo learning loop that extracts patterns into the canonical knowledge base.