Architect-Refine-Critique

github:NTCoding/claude-skillz · first committed 2026-01-18 · source · badge

Three-phase design review. Chain architect â refiner â critique subagents.

Listed, not yet scored. This skill has no public model test yet. It is in the scoring queue; submitted skills go first, then the most-installed.

15 commits to this file on GitHub (newest first).

2026-01-31 · commit f9e68f5 · Fix skill frontmatter names: use valid slugs matching directory names

2026-01-29 · commit 3eaac69 · Add skill activation enforcement to launcher and trigger descriptions to skills

2026-01-18 · commit 2c52e9d · test

2026-01-18 · commit b8cc10c · Revert architect-refine-critique SKILL.md to version before 34cf1ad

2026-01-18 · commit 34cf1ad · revert to using skills

2026-01-18 · commit 2b65715 · Refactor /arc to use chained subagents with ultra-thin orchestrator

2026-01-18 · commit 4e2586f · Remove all instances of 'plan' from architect-refine-critique

2026-01-18 · commit 6eca29f · Fix Architect description - no constraints

2026-01-18 · commit d453622 · Specify exact Task tool syntax for sub-agent spawning

2026-01-18 · commit 23a53bb · Broaden Critique phase: wrong, better, simpler

2026-01-18 · commit c6ce8d1 · Fix architect-refine-critique: explicit VERIFY steps, remove over-specification

2026-01-18 · commit 6bf65e5 · Move Invoking section before checklist so [name] and [target] are defined first

2026-01-18 · commit 5ba1f5b · Add separation-of-concerns to Refiner phase

2026-01-18 · commit 300d530 · Fix architect-refine-critique to run phases sequentially

2026-01-18 · commit 64c7dd0 · Add tactical-ddd skill and architect-refine-critique workflow