rawsql-ts

Assess and dispatch rawsql-ts developer tasks by tracing repository impact, concept and package boundaries, verification cost, difficulty, risk, model and reasoning effort, and Codex worktree requirements. Use when a new rawsql-ts issue, bug, feature, refactor, investigation, CI failure, review request, migration, or documentation change must be analyzed before action or handed to a new Codex task.

Report per-item attainment for completed rawsql-ts developer work.

Review tracked rawsql-ts package-level Scope, Test Policy, Authority Model, Technology Policy, review-plan, and generated review views before package implementation or documentation changes. Use only when those package-owned artifacts exist in the checked-out base.

Prepare rawsql-ts pull requests with the repository PR template and local readiness gate so required merge, CLI migration, and scaffold proof sections are not missed.

Check tmp/RETRO.md before PR or human review so unresolved retro items do not slip through rawsql-ts developer workflow.

Capture meaningful recognition mismatches, false completion claims, and verification misses into tmp/RETRO.md for rawsql-ts developer work.

Run two-cycle self-review and triage before presenting rawsql-ts developer work to a human reviewer or requester.

Review public and transformer API output shape decisions for SQL/AST/model transformations, ensuring callers can choose model versus string output and customize formatting when string output is produced.

Classify and review rawsql-ts Changeset entries for patch, minor, major, no-release, stale entries, public API additions, bug fixes, docs/demo updates, and reviewer comments that challenge release-note wording or bump level.

Implement or review rawsql-ts packages/core parser, analyzer, formatter, SQL syntax, AST, SelectValueCollector, SelectOutputCollector, or SelectBodyExtractor changes with DBMS-neutral boundary checks, regression tests, benchmark/demo gates, and release-note classification.

Manage rawsql-ts external review-tool usage for CodeRabbit or similar AI review bots when PRs are stacked, small, generated, broad, rate-limited, or likely to consume review quota before the branch is ready.

Prepare scoped review packets for broad rawsql-ts diffs with generated API docs, generated review pages, mass removals, or many package/docs changes. Use when a PR is likely to exceed reviewer or review-tool limits, when generated files dominate the diff, or when reviewers need a focused map from source changes to generated artifacts and verification gates.

Review rawsql-ts structured metadata migrations and generators for schema/version parity, real fixture parsing, canonical enum alignment, relationship evidence integrity, display-label quality, and generated review-view consistency. Use when work adds or migrates Concept Specs, Scope/Test/Authority/Technology rule registries, AI review JSON, relationship metadata, or generated review docs.

Draft acceptance items and verification methods for rawsql-ts developer work.

Review upstream ConceptSpec drafts, issue notes promoted toward ConceptSpec, or early design memos before they are treated as stable grounding.

Organize early hearing notes, customer interview notes, rough memos, and chaotic domain discussions into ConceptSpec intake buckets without finalizing concepts or creating implementation tasks.

Check business-bearing or domain-meaning issues before implementation planning to confirm grounding in Concept Specs, DFDs, Process Maps, Package Scope, feature-local specs, or explicit technical-support classification.