jawadsajjad503/workbuddy-skill/skill-progressive-refactor
Claude Code and agent skill by jawadsajjad503/workbuddy-skill
- community
Checking that the SKILL.md file is still in the repository.
Use when a SKILL.md is too long or costly to load and needs a safe progressive-disclosure refactor into focused references without changing its contract.
Install
npx vskill@latest install jawadsajjad503/workbuddy-skill/skill-progressive-refactorSKILL.md
The same SKILL.md was published earlier as sandbaseai/workbuddy-skill/skill-progressive-refactor.
Skill Progressive Refactor
Use this Skill after a quality audit identifies excessive length, duplicated detail, or poor progressive disclosure. It changes files only within the authorized Skill directory, preserves the original revision for rollback, and must not silently alter the Skill's trigger, safety policy, output contract, or frontmatter identity.
Freeze the contract
Read the target SKILL.md completely and inventory its content hash, line and character count, frontmatter, triggers, tools, commands, outputs, stopping conditions, links, and existing references/. Read repository instructions and native validators. Record the baseline and make a bounded refactor plan; for this authorized maintenance workflow, proceed without an extra approval round unless a scope, ownership, or destructive-action ambiguity exists.
Separate content into:
- core control flow: triggers, phases, decisions, safety boundaries,
prerequisites, and output handoff; keep these in SKILL.md;
- on-demand detail: help, templates, examples, reference tables, domain
knowledge, long checklists, and explanations; move these to single-purpose files under references/.
Do not split a sentence so that its guard, exception, or stopping condition is left behind. Preserve exact names, stable anchors, commands, and cross-links; never rename a colon-form name or external identifier merely to satisfy a style preference.
Perform the smallest safe extraction
Create or update one focused reference file per coherent topic. Each file gets a clear heading, purpose, prerequisites, and an explicit trigger sentence in the core. Replace extracted blocks with concise pointers such as “Read the relevant reference file when …”. Keep references bounded and avoid creating a second hidden router or duplicating the same rule in multiple locations.
The compact core should make the full workflow understandable quickly. A target of 100 lines is a useful warning threshold, not permission to delete necessary safety or contract material. Preserve tables/examples when their structure carries meaning; compress prose only after checking that the result still entails the original behavior.
Validate behavior and disclosure
Run repository-native frontmatter, link, path, markdown, package, and relevant tests. Confirm that the core is within the agreed budget, every reference is reachable from a documented trigger, all expected outputs remain reachable, and no old links or commands point at removed content. Inspect the diff for lost negations, weakened authorization, changed paths, secret exposure, accidental network/command expansion, or references loaded unconditionally.
Re-run the complete Skill quality audit, including security, license, capability, description-length, and model/dependency checks. Compare before and after behavior on representative trigger prompts or deterministic fixtures; record tests not run and environment limits. If a check fails, fix the bounded refactor or restore the affected file from the recorded revision rather than publishing a partial split.
Report and rollback
Return the original and final hashes, line/character counts, files created or changed, extraction map, contract checks, test commands and exit codes, quality-audit result, unresolved limitations, and rollback revision. Preserve the original content in version control; do not delete history or unrelated references. End only after the next audit command and the maintainer's next review trigger are explicit.
Security
- Security scanPassed, 95 of 100Pattern checks on the skill's files.
- Model reviewPassed
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingskills/skill-progressive-refactor/SKILL.md
Popularity
- 7-day trend score
- 5
Not refreshed yet.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/jawadsajjad503/workbuddy-skill/skill-progressive-refactor)