dazzle

Use when authoring trial.toml, running agent QA ladder (coverage/journey/deep), or triaging friction for improve auto-seed. Agent-first live investigation of Dazzle apps; human is gated L4. Triggers: qa trial, trial.toml, business-user testing, coverage inventory, journey mode, adoption_criteria, ownership triage, #1625 ladder.

Use when authoring trial.toml, running agent QA ladder (coverage/journey/deep), or triaging friction for improve auto-seed. Agent-first live investigation of Dazzle apps; human is gated L4. Triggers: qa trial, trial.toml, business-user testing, coverage inventory, journey mode, adoption_criteria, ownership triage, #1625 ladder.

Reconstruct Dazzle/HM judgement from stems before inventing structure

Run all quality checks on modified files — the on-demand pre-ship quality gate

Monitor CI pipeline status, diagnose failures, and drive the badge back to green

Commit, verify, tag, and push with the repo's pre-flight gate suite (lint, type, drift gates, docs build)

Reconstruct Dazzle/HM judgement from stems before inventing structure

Use when writing or editing Dazzle .dsl files — syntax rules, the mistakes agents actually make, and pointers to the drift-gated references. Complements (never replaces) CLAUDE.md's DSL Quick Reference.

Use when the user wants to generate a stakeholder-facing English specification document (SPECIFICATION.md) from a Dazzle DSL project — for investors, business leaders, founders, or non-technical decision-makers. Reverses the usual DSL→app flow into DSL→prose.

Use when writing or editing Dazzle .dsl files — syntax rules, the mistakes agents actually make, and pointers to the drift-gated references. Complements (never replaces) AGENTS.md's DSL Quick Reference.

Use when the user wants to generate a stakeholder-facing English specification document (SPECIFICATION.md) from a Dazzle DSL project — for investors, business leaders, founders, or non-technical decision-makers. Reverses the usual DSL→app flow into DSL→prose.

Bump the semantic version across all six canonical locations and update CHANGELOG

Sync documentation with recently closed GitHub issues

Two-phase read-only code-smells analysis: regression checks + new systemic patterns

Use when executing a multi-phase task autonomously with advance authority to advance between phases — turns a list of phases into a gate-driven loop where each phase completes only when its machine-checkable gate exits 0. Use when the user grants autonomy ("just keep going", "don't stop to ask", "work through the whole list", "max effort", token-rich), or hands you a phased plan to execute end-to-end. Maintains PLAN.md, auto-proceeds on green, escalates only on the fixed list below.