agentleFS
aws-solutionsaws-solutions/konductor/skills/constraints/SKILL.md

constraints

Skill11 starsChanged 49 days agoApache-2.0View on GitHub

MUST USE for every interaction across all Konductor agents. Defines the maker-checker, quality gate, severity ordering, and failure recovery rules; the always-loaded core rules hold the rest.

4 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

3 rules. 0 are common, 3 are distinctive, and 0 are very specific to this project.

  • 0

    Common

    Most of the 4 edited versions we read keep these.

  • 3

    Distinctive

    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:

    After 2 failed fix attempts for the same error, STOP.

    line 39
  2. Distinctive:

    Only then propose a fix

    line 43
  3. Distinctive:

    ALWAYS DO rule not followed

    line 64

“Common” is measured: kept by at least 6 in 10 of the 4 edited versions of this file we could read. “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

Constraints

Overview

Hard rules that ALL Konductor agents must follow in every interaction. These are non-negotiable — violations are treated as CRITICAL failures regardless of context.

Usage

This skill is always active. It applies to every task across all Konductor agents. No explicit activation needed.

Core Rules

NEVER DO

The NEVER DO rules live in the core rules, which every agent loads at session start.

ALWAYS DO

  1. Read before edit: see Engineering practices in the core rules
  2. Follow existing patterns: see Engineering practices and Writing habits in the core rules
  3. Maker-checker — Always run the maker-checker pattern: generate artifact → validate with the corresponding checker skill before presenting to the user
  4. Backward compatibility: see Engineering practices in the core rules
  5. Comment complex logic: see Engineering practices in the core rules
  6. Structured logging: see Engineering practices in the core rules
  7. Verify before done: see "No broken commits" and "No broken state" under NEVER DO, and Engineering practices, in the core rules
  8. Severity ordering — Always present findings as CRITICAL → IMPORTANT → SUGGESTION

Failure Recovery

2-Strike Circuit Breaker

After 2 failed fix attempts for the same error, STOP. Do not attempt a third fix. Instead:

  1. Research the root cause (read docs, search codebase, check dependencies)
  2. Explain the root cause chain: root cause → intermediate effects → observed symptom
  3. Only then propose a fix

Unknown Systems

See Engineering practices in the core rules: research first, fix second.

Acknowledge Uncertainty

See Engineering practices in the core rules: for an unverified claim that matters, say whether it rests on documented knowledge, inference, or a guess.

Quality Gate

CRITICAL (immediate stop):

  • Any NEVER DO rule violated (see the core rules)
  • Artifact presented to user without checker validation (maker-checker bypass)
  • Code left in non-compiling state
  • Secrets or credentials in source code

IMPORTANT (must fix before proceeding):

  • ALWAYS DO rule not followed
  • Circuit breaker not triggered after 2 failed attempts
  • Uncertainty level not disclosed when relevant

SUGGESTION:

  • Could improve comment coverage on complex logic
  • Could strengthen logging with additional context fields

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.