agentleFS
Ezra144israelEzra144israel/agent-teamwork-kit/skills/better-coding/SKILL.md

better-coding

Skill24 starsChanged 56 days agoMITView on GitHub

Evaluate minimum-sufficient implementation routes and enforce the selected route for a fixed, authorized outcome. Use after outcome, scope, authority, and acceptance evidence are fixed. Do not choose outcomes, set acceptance, adjudicate governance, grade finished work, design tests alone, conduct security audits, or perform unrelated cleanup.

1 other project uses a version of this.

Adopted by

Most recently taken up by

  • Other versions1 project uses a version of this.

    Word for word · 1

    1. Ezra144israel/governed-agent-skills★ 24100% same

    Edited · 0

    None found.

    Wrote their own version · 0

    None found.

    How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.

  • In this repository7 other files it gives its agents.

The file

We split this skill into the rules it gives an agent

7 rules. About 0 are common, 7 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈7

    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:

    Use the fewest concepts that satisfy the approved result and its acceptance evidence.

    line 8
  2. Distinctive:

    If outcome, acceptance evidence, or authority is unresolved, return to its owner.

    line 20
  3. Distinctive:

    Keep this evidence short. That owner decides when a proposal is required and how it returns.

    line 51
  4. Distinctive:

    When routes satisfy different checks and the current evidence does not identify the applicable check, return that gap to whoever owns the …

    line 54
  5. Distinctive:

    Treat file size as a signal, not a diagnosis.

    line 67
  6. Distinctive:

    Before editing, inspect the repository status and the complete target diff.

    line 77
  7. Distinctive:

    Stop when the selected acceptance evidence passes.

    line 88

≈ “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

Better Coding

Use the fewest concepts that satisfy the approved result and its acceptance evidence. Minimize maintenance burden, not raw line count.

Keep the ownership boundary

  • thinking owns the stage, outcome, scope, evidence, route recommendation, hard stops, and the transition to execution.
  • teamwork owns authority, seats, formal contracts, verdicts, publication, and Done.
  • This skill owns route comparison, ownership seams, concept control, readability, and implementation verification.

If outcome, acceptance evidence, or authority is unresolved, return to its owner. Apply this lens only to the selected route and safe slice. A finished change needs independent review.

Compare routes before adding code

  1. Restate the fixed result and acceptance evidence without reopening them.
  2. Inspect current code and reuse candidates. Compare current behavior, configuration, reuse, adaptation, replacement, new code, deletion, documentation, and no-code routes. If the current evidence already passes, make no change.
  3. Compare the strongest simpler alternative. Choose a route only when the fixed acceptance evidence can prove it.
  4. Name the smallest ownership seam and the concepts the selected route needs. Give each new responsibility a concrete reason tied to the accepted outcome. Name tempting concepts the result does not need.

Choose reuse when it is the better total solution. Existing code alone is no reason to require reuse or reject new code. Compare simplicity, safety, clarity, coupling, and maintenance cost. Existing tools, frameworks, vendor implementations, standards, protocols, and delivery systems are options. Building a capability, connection, protocol, or delivery mechanism remains valid when it is the better lawful solution. Neither reuse nor custom work wins without evidence.

For an implementation return or an engineering proposal required by teamwork, name the inspected source and reuse candidates, what to reuse, adapt, replace, or build and why, the strongest simpler alternative, necessary new responsibilities, material placement and dependency choices, verification and meaningful negative cases, and unresolved prerequisites or decisions that could change implementation. Keep this evidence short. That owner decides when a proposal is required and how it returns. This skill creates no proposal or approval gate.

When routes satisfy different checks and the current evidence does not identify the applicable check, return that gap to whoever owns the outcome and acceptance evidence. Do not create a second implementation brief or return schema.

Bound the implementation

Decline a dependency, layer, abstraction, state, endpoint, migration, generic helper, or compatibility path whose only reason is hypothetical reuse or an unselected future. Prefer readable, explicit, testable code over dense code. A longer implementation is smaller when it carries fewer concepts or less maintenance risk.

Treat file size as a signal, not a diagnosis. Extract a bounded seam only when the extraction lowers verified current risk. Remove only code made obsolete by the selected change. Report unrelated cleanup instead of doing it.

For external calls, choose the failure path deliberately: timeout, retry, idempotency, partial failure, and secret handling where they apply. Load project security or infrastructure detail for those boundaries.

Verification and repair discipline

Before editing, inspect the repository status and the complete target diff. Separate defects introduced by the selected change from pre-existing defects. Pre-existing issues stay out of scope unless the operator expands the contract.

Match verification to the fixed acceptance evidence. Before returning, inspect the final diff and status again, then run the required repository gates. Advisory or judgment-heavy measurement must not become a mandatory deployment gate. After each repair, run the narrowest relevant focused recheck. Never weaken assertions, remove coverage, suppress errors, or change expected behavior to obtain a green result.

Stop when the selected acceptance evidence passes. Do not add cleanup or future-proofing after that point.

For tasks that need mechanical containment, the optional Enforcement Layer change-guardian can seal allowed change classes and bind verification to the exact final repository state. It does not replace this Instruction Layer skill's maintainability judgment or independent review.

With teamwork active, return the route, seam, declined concepts, verification, and residual risk in its Builder judgment record.

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.