agentsclimarketplace

Spring refresh

Skill joaquimscosta/arkhe-claude-plugins/plugins/spring-boot/skills/spring-refresh

Check Spring Boot skill content freshness against latest research and flag skills needing updates. Use when running /spring-refresh, or when user mentions "refresh spring skills", "spring boot update check", "skill freshness", "stale spring content", "update spring boot skills".From its SKILL.md

Install
npx -y skills add joaquimscosta/arkhe-claude-plugins --skill spring-refresh

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 21 stars21 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

3.6 KB, 806 tokens by cl100k_base, as published. Nobody here has run it

Spring Boot Skill Freshness

Check whether Spring Boot skills are current with the latest research docs.

Arguments

Parse $ARGUMENTS for mode:

  • check (default, no args) — Scan and report freshness
  • refresh — Update research docs via deep-research, then report
  • update <skill-name> — Help update a specific skill's content

Mode: check

  1. Run the freshness scanner:
python3 ${CLAUDE_SKILL_DIR}/scripts/scan_skill_freshness.py --format json ${CLAUDE_SKILL_DIR}/../..
  1. Parse the JSON output and present a formatted report with:

    • Research doc staleness table (file, version, last updated, age, status)
    • Skill drift table (name, target version, last modified, research date, drift, status)
    • Summary counts (up-to-date, needs review, needs update)
  2. For skills with NEEDS_REVIEW or NEEDS_UPDATE:

    • List which research docs they depend on
    • Suggest running /spring-refresh update <skill-name> for each

Mode: refresh

  1. For each Spring Boot research doc, invoke the Skill tool with core:deep-research to refresh:

    • refresh spring-boot-ecosystem (for ecosystem-research.md)
    • refresh spring-boot-ddd-implementation (for ddd-implementation.md)
    • refresh spring-boot-security-observability-testing (for security-observability-testing.md)
  2. After all research docs are refreshed, run the check mode to produce the updated report.

  3. Show a before/after comparison of research doc dates.

Fallback: If the deep-research skill is not available, skip the research refresh and run check mode only. Inform the user they can manually update research docs.

Mode: update <skill-name>

  1. Run check mode first to identify drift status for the specified skill.

  2. Read the research docs that map to this skill:

    • spring-boot-ecosystem-research.md → scanner, verify, web-api, data-ddd, modulith, domain-driven-design
    • spring-boot-ddd-implementation.md → domain-driven-design, data-ddd, web-api, modulith
    • spring-boot-security-observability-testing.md → security, observability, testing
  3. Read the skill's current content: SKILL.md, EXAMPLES.md, TROUBLESHOOTING.md, and all files in references/.

  4. Compare research content against skill content. Identify:

    • New APIs or patterns in research not covered in the skill
    • Deprecated patterns in the skill that research has replaced
    • Version-specific changes (e.g., new Spring Boot minor version features)
  5. Present proposed changes to the user for confirmation before editing.

  6. After updates, bump the spring-boot-version field if targeting a new version.

Output Format

## Spring Boot Skill Freshness Report

### Research Documents
| Document | Version | Updated | Age | Status |
|----------|---------|---------|-----|--------|

### Skills
| Skill | Target | Modified | Research | Drift | Status |
|-------|--------|----------|----------|-------|--------|

### Recommendations
- [actionable items]

References

What ships with it: 3 files

20.4 KB alongside SKILL.md, 1 of them executable

scripts/

Keep looking

Skills are one crate of 325,949. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.