Skip to content

JPeetz/agent-skills / skills/development/finding-unknowns/skills/progressive-disclosure

progressive-disclosure

Verified

Use when splitting oversized skills into on-demand files.

FileKindSize
CHANGELOG.mdmarkdown409 B
SKILL.mdskill md2.7 KB
evals/evals.jsonconfig2.7 KB

evals/evals.json

View on GitHub
{
  "skill": "progressive-disclosure",
  "version": "1.0.0",
  "cases": [
    {
      "id": "e1",
      "description": "Should trigger: oversized skill paying for unused content every turn",
      "input": "This skill is 400 lines but most runs only need the first 30. Restructure it.",
      "shouldTrigger": true,
      "expect": "Splits branches into sibling files with condition-bearing pointers",
      "assertions": [
        "Identifies branches (genuinely different run paths)",
        "Sorts sections into universal core vs branch-specific",
        "Keeps universal core in entry file, still coherent alone",
        "Names sibling files for what they hold (verification.md, glossary.md)",
        "Writes pointers naming condition AND file together",
        "Walks each branch end-to-end, reports new entry size"
      ]
    },
    {
      "id": "e2",
      "description": "Should trigger: moving on-demand material behind a pointer",
      "input": "Move the verification section out of the main skill file so it only loads when the user asks to verify.",
      "shouldTrigger": true,
      "expect": "Relocates the verification branch to verification.md with a conditional pointer",
      "assertions": [
        "Pointer states the triggering condition",
        "Entry file still reads coherently"
      ]
    },
    {
      "id": "n1",
      "description": "Not trigger: deleting dead material instead of splitting",
      "input": "Remove the outdated glossary section entirely, nobody uses it.",
      "shouldTrigger": false,
      "expect": "Deletes the section outright (splitting is not deleting)",
      "assertions": [
        "No sibling file created; section removed"
      ]
    },
    {
      "id": "n2",
      "description": "Not trigger: audit of contradictions, not restructuring",
      "input": "Do a context audit on my skill files.",
      "shouldTrigger": false,
      "expect": "Routes to context-audit workflow",
      "assertions": [
        "No file splitting performed here"
      ]
    },
    {
      "id": "e3",
      "description": "Should trigger: splitting a glossary into an on-demand file",
      "input": "The glossary section is large and only needed when terminology is unfamiliar. Move it to its own file with a clear pointer.",
      "shouldTrigger": true,
      "expect": "Moves glossary to glossary.md with a pointer naming the condition and file together",
      "assertions": [
        "Names the sibling file for what it holds (glossary.md)",
        "Writes a pointer that names the condition AND the file together",
        "Entry file still reads coherently on its own",
        "Walks the branch to confirm it has everything it needs"
      ]
    }
  ]
}
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: 472kUpdated 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