Skip to content

Releases: sungjunlee/craftkit

v0.1.1 — clearer skill names

13 Apr 06:27
@sungjunlee sungjunlee
v0.1.1
5bff7e7

Choose a tag to compare

View all tags
v0.1.1 — clearer skill names

Changed

Renamed three skills for clearer invocation semantics and to end a root-word collision:

  • craft-reflect → craft-critique — matches what users actually type ("critique this," "review this") instead of the softer, more ambiguous "reflect."
  • craft-research → craft-survey — ends the root-word collision with craft-autoresearch, which does eval-driven optimization (not prior-art research). Distinct verbs now signal distinct skills.
  • craft-blueprint → craft-scaffold — matches the established programming "scaffolding" metaphor (Rails, Yeoman, etc.).

In-body terminology in the renamed skills was updated to match. Output-format rules and eval-driven scoring behavior are unchanged. All cross-references (README, AGENTS.md, docs/, sibling skills' references) were updated in the same commit.

Migration

No compat shim. The renamed directories and frontmatter name: fields are the authoritative trigger mechanism. The description: fields still include the old trigger words (e.g. "blueprint," "reflect," "research") as aliases, so natural-language invocation keeps working.

Full changelog

See CHANGELOG.md.

Loading

v0.1.0 — initial public-ready release

13 Apr 06:29
@sungjunlee sungjunlee
v0.1.0
496bb1d

Choose a tag to compare

View all tags
v0.1.0 — initial public-ready release Latest
Latest

Initial public-ready release. Six craft-* skills for prompt and skill authoring, all optimized through craft-autoresearch eval-driven passes against scored rubrics.

Skills

Skill Use when
craft-prompt a new prompt is needed from scratch for any LLM
craft-blueprint a rough idea needs structure before implementation
craft-reflect a prompt or skill "feels off" and needs a diagnostic pass
craft-tune an existing prompt needs targeted, minimal-diff sharpening
craft-research a new skill should be grounded in prior art
craft-autoresearch a prompt or skill works "sometimes" and needs eval-driven iteration

Note: In v0.1.1 three of these skills were renamed for clarity — craft-reflect → craft-critique, craft-research → craft-survey, craft-blueprint → craft-scaffold. See the v0.1.1 release notes.

Install

  • npx skills add sungjunlee/craftkit
  • /plugin marketplace add https://github.com/sungjunlee/craftkit.git → /plugin install craftkit@craftkit

Full changelog

See CHANGELOG.md.

Assets 2
Loading