agentleFS
renanflustosarenanflustosa/octo-cluster/.claude/skills/ponytail-lite/SKILL.md

ponytail-lite

Skill3 starsChanged 23 days agoMITView on GitHub

Minimal implementation ladder before writing code. Apply before each edit; question over-scope during planning. Complements caveman (prose). Inspired by DietrichGebert/ponytail (MIT).

1 other project uses a version of this.

Adopted by

Most recently taken up by

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:

    Before writing any code, stop at the first rung that holds:

    line 14
  2. Distinctive:

    Order: read and trace the real flow first, then climb.

    line 24

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 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

Ponytail-lite — minimal implementation

Adapted from DietrichGebert/ponytail (MIT).

Lazy means efficient, not careless. Complements caveman-mode rule (shorter replies). This skill governs what to implement, not how to phrase status updates.

The ladder

Before writing any code, stop at the first rung that holds:

  1. YAGNI — Does this need to be built at all?
  2. Reuse — Does it already exist in this codebase? Reuse the helper, util, or pattern — do not rewrite.
  3. Stdlib — Does the standard library already do this?
  4. Native — Does a native platform feature cover it? (<input type="date">, Intl, path.Join, etc.)
  5. Installed dep — Does an already-installed dependency solve it?
  6. One line — Can this be one line? Make it one line.
  7. Minimum — Only then: write the minimum code that works.

Order: read and trace the real flow first, then climb. Lazy about the solution, never about reading.

Examples by rung

RungBadGood
YAGNIBuild a generic plugin system for one hookInline the one hook
ReuseNew parseDate() when utils/date.ts existsImport existing helper
StdlibHand-rolled email regex class"@" in email or stdlib validator
Nativeflatpickr for a date field<input type="date">
Installed depAdd lodash for one groupByUse existing dep or stdlib
One line15-line wrapper around dict(zip(...))dict(zip(keys, values))
MinimumFramework for a one-off scriptDirect script

Bug fix = root cause

A report names a symptom. Grep every caller of the function you touch and fix the shared function once — one guard there beats one patch per caller.

Carve-outs (never cut)

  • Understanding the problem (read fully, trace flow before picking a rung)
  • Input validation at trust boundaries
  • Error handling that prevents data loss
  • Security, accessibility
  • Hardware calibration (clocks drift, sensors read off)
  • Anything explicitly requested by the user or ticket

Non-trivial logic leaves ONE runnable check (assert-based self-check or one small test file; no frameworks, no fixtures). Trivial one-liners need no test.

Debt comments

Mark intentional simplifications:

# ponytail: global lock; upgrade to per-key sharding if contention shows up

Names the ceiling and the upgrade path so "later" does not become "never".

Phase integration

PhaseBehavior
PlanRejected alternatives cite rung; plan must not violate carve-outs
ExecuteLadder active before each file edit
/shipPrefer smallest diff that passes the boundary gate

Does NOT replace

  • caveman-mode rule — prose compression on status turns
  • systematic-debugging — reproduce before fix; ponytail-lite applies after root cause is understood

Trigger: active when .cursor/rules/ponytail-lite.mdc applies (Cursor rule; imported by CLAUDE.md in Claude Code).

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.