LLM-judgment critique of documentation quality — the ceiling counterpart to the rule-based doc floor (detect-doc-drift / check-docs / docs-pipeline, which enforce existence, link freshness, coverage). Asks whether a doc teaches, whether the order matches the reader's mental model, whether examples earn their place…
Conductor for the fleet family — one tier above the members, coordinating the fleets themselves rather than fanning out over an item queue. Probes each installed fleet's queue through that member's gate-free path only, derives the run as a hybrid dependency DAG (a CI trust gate first, then ideation, then intake with…
Check a knowledge graph for content that cannot be trusted — connectors that stamped a fresh sync timestamp while hard-failing, and extractor-derived nodes minted out of prose rather than code. Orchestrates harness graph integrity; never reimplements detection. Reports denominators, so a run that inspected nothing…
Strategy-grounded ideation fan-out at the head of the fleet spine — derive a queue of disjoint themes from STRATEGY.md tracks and supplied opportunity areas, confirm the batch once, fan out worktree-isolated subagents that each run the real harness-ideate pipeline for one theme to its ranked artifact, collect every…
Pre-brainstorm ideation phase. Generates N candidate ideas grounded in STRATEGY.md (when present), critiques each against its strongest objection, ranks by (impact × confidence) ÷ effort with a bounded strategy-alignment tiebreaker, and writes a single ranked Markdown artifact to docs/ideation/[slug]-YYYY-MM-DD.md.…
Scaffold or migrate a test-suite project (API, E2E/UI, or shared library) with test-suite-specific layer models, tags, reporter stack, and custom report.