agentleFS
YehudaFrankelYehudaFrankel/clankbrain/.claude/skills/shadow-code/SKILL.md

shadow-code

Skill16 starsChanged 7 months agoMITView on GitHub

Developer sharpening exercise — user writes the implementation first, Claude reviews and compares. Keeps independent coding ability sharp. Triggers on "/shadow-code", "shadow mode", "I'll write it first", "let me try first".

The file

We split this skill into the rules it gives an agent

2 rules. About 0 are common, 2 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Never be dismissive of a working approach that differs from Claude's style

    line 44
  2. Distinctive:

    If the user's approach is valid but unconventional, say so explicitly — don't imply it's wrong

    line 45

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Skill: shadow-code

Trigger: /shadow-code or "shadow mode" or "I'll write it first" or "let me try first"

Description: You write the implementation. Claude reviews it, compares to what it would have written, and explains any differences — without rewriting unless you ask. Keeps the coding muscle sharp.


Steps

  1. Confirm the task — restate in one sentence what needs to be implemented. Make sure scope is clear before the user starts.
  2. Step back — do not write any implementation code. Provide only:
    • The relevant file path and line number to work from
    • Any project-specific conventions that apply
    • What the method/function signature should be
  3. Wait for the user's implementation — they paste or describe what they wrote.
  4. Review without rewriting — compare their implementation to what Claude would have written. Report:
    • Correct: what they got right
    • Gaps: anything missing or wrong, with explanation of WHY (not just what)
    • Differences: where Claude would have done it differently, and whether their approach is also valid or strictly worse
    • Project gotchas: any platform-specific pattern they may not have applied
  5. Score it honestly:
    • ✅ Production-ready — would ship as-is
    • 🟡 Mostly right — minor fixes needed (list them)
    • 🔴 Needs rework — fundamental issue (explain clearly, then offer to show the correct version)
  6. Only rewrite if asked — if the user says "show me yours" or "fix it", then Claude writes the corrected version. Otherwise stop at the review.

Notes

  • This skill is about building the muscle, not shipping fast — be thorough in the review
  • Never be dismissive of a working approach that differs from Claude's style
  • If the user's approach is valid but unconventional, say so explicitly — don't imply it's wrong
  • The goal is the gap analysis, not the grade
  • Suggested cadence: every 5-10 sessions, or when a feature is small enough to attempt solo

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.