refineprompt

name: RefinePrompt description: "Apply targeted transforms to prompt-shaped documents: align conventions, debrand vendor references, minimize filler, rescope tool grants, extract bulk reference, adapt for downstream repos. USE WHEN refining an adopted skill, polishing an authored skill/rule/agent, removing rot, or porting prompts across repos." version: 0.1.0 allowed-tools: Read, Edit, Write, Grep argument-hint: "[target file or repo]"

Six transforms for refining a prompt-shaped document. Apply each independently or in sequence; each is also composable by AdoptArtifact during community-skill adoption.

Decide what to apply

Transform	Apply when
Align	Indent, fence tags, heading depth, frontmatter, or table alignment is off
Debrand	Hardcoded vendor names or external-service dependencies bias the artifact
Minimize	Motivational, marketing, or filler prose; emphasis inflation
Rescope	`allowed-tools: '*'` or unnecessarily broad grants
Extract	SKILL.md body has bulk reference material that does not need to be loaded
Adapt	Porting prompts from this repo to a downstream/consumer repo

Transforms

@Align.md

@Debrand.md

@Minimize.md

@Rescope.md

@Extract.md

@Adapt.md

Cross-cutting constraints

Never rewrite content to change meaning. Refine structure, remove rot, narrow scope; do not paraphrase.
Preserve anti-pattern lists, concrete examples, command/regex/schema blocks, and runtime-critical frontmatter (name, description, version, allowed-tools, argument-hint, hooks).
Recompute the artifact's SHA-256 after every transform and sync it to the provenance sidecar if one exists. The sidecar's subject.digest.sha256 must match the current file content.
When in doubt, preserve and flag for human review rather than transform silently.