Skip to content

JPeetz/agent-skills / skills/development/joerg-instruction-execution

joerg-instruction-execution

Verified

Use this skill when executing Joerg's (or another profile's) shorthand instructions — "try it", "Ok", "wire it, of course", or a correction. Capture intent from his style, prompt properly from it, and verify on the output side before claiming done.

FileKindSize
CHANGELOG.mdmarkdown460 B
SKILL.mdskill md7.6 KB
evals/evals.jsonconfig2.2 KB

evals/evals.json

View on GitHub
{
  "skill": "joerg-instruction-execution",
  "version": "1.0.0",
  "cases": [
    {
      "id": "e1",
      "description": "Should trigger: terse shorthand instruction",
      "input": "try it",
      "shouldTrigger": true,
      "expect": "Interprets intent and executes without a confirmation question",
      "assertions": [
        "Does not ask for permission on the obvious default",
        "Delivers output and verifies on the output side",
        "Reports honestly with evidence if blocked"
      ]
    },
    {
      "id": "e2",
      "description": "Should trigger: correction loop as authoritative feedback",
      "input": "What the heck is this, the lip sync is still awful!",
      "shouldTrigger": true,
      "expect": "Treats the correction as the real spec and re-derives the approach",
      "assertions": [
        "Does not argue with the feedback",
        "Changes the output until it matches",
        "Logs the lesson to the relevant skill"
      ]
    },
    {
      "id": "e3",
      "description": "Should trigger: ambiguous scope where ask-first is warranted",
      "input": "Add a search feature to the dashboard. You decide the rest.",
      "shouldTrigger": true,
      "expect": "Resolves via context layers then asks only outcome-changing questions",
      "assertions": [
        "Checks JEV -> MeMex -> HermesVault -> disk before asking",
        "Asks 1-5 structured questions with options, recommended first",
        "Surfaces assumptions explicitly if proceeding"
      ]
    },
    {
      "id": "n1",
      "description": "Not trigger: fully-specified mechanical task (ACT-DON'T-ASK)",
      "input": "Rename the variable 'usreName' to 'userName' in src/auth.js.",
      "shouldTrigger": false,
      "expect": "Executes mechanically without the intent/verification loop",
      "assertions": [
        "No scope questions needed",
        "Direct read-file/patch/verify execution"
      ]
    },
    {
      "id": "n2",
      "description": "Not trigger: general factual question",
      "input": "What is the capital of France?",
      "shouldTrigger": false,
      "expect": "Answers directly, no instruction-execution protocol",
      "assertions": [
        "No verification checklist needed"
      ]
    }
  ]
}
reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis.Stars: 471kUpdated Claude CodeCodexVerifiedInstall
diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness.Stars: 322kUpdated Claude CodeCodexVerifiedInstall
dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependenciesStars: 322kUpdated Claude CodeCodexVerifiedInstall
executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is availableStars: 322kUpdated Claude CodeCodexVerifiedInstall