Skip to content

JPeetz/agent-skills / skills/development/finding-unknowns/skills/change-quiz

change-quiz

Verified

Use when reviewing what changed after a long session before merging — build a change report grouped by intent, then quiz the user so unread changes don't ship.

FileKindSize
CHANGELOG.mdmarkdown368 B
SKILL.mdskill md2.4 KB
evals/evals.jsonconfig2.4 KB

evals/evals.json

View on GitHub
{
  "skill": "change-quiz",
  "version": "1.0.0",
  "cases": [
    {
      "id": "e1",
      "description": "Should trigger: post-session review before merge",
      "input": "What did we actually change this session? Quiz me before I merge.",
      "shouldTrigger": true,
      "expect": "Builds a 4-section change report then quizzes 5-8 questions, graded honestly",
      "assertions": [
        "Groups 'what changed' by intent (feature/fix/refactor), not by file",
        "Covers how the change interacts with existing code paths",
        "Asks 5-8 questions targeting deviations, edge cases, interactions",
        "Grades honestly: critical misses = fail even at 5/8",
        "Explains each miss and flags gap vs too-clever"
      ]
    },
    {
      "id": "e2",
      "description": "Should trigger: verification of understanding before shipping",
      "input": "I'm about to merge the deploy changes but I want to make sure I understand the retry behavior now.",
      "shouldTrigger": true,
      "expect": "Produces report + quiz weighted toward interaction effects",
      "assertions": [
        "Includes 'Intuition' mental-model updates",
        "Mix of recall and prediction questions"
      ]
    },
    {
      "id": "n1",
      "description": "Not trigger: code review of someone else's diff",
      "input": "Review this PR from our intern for style issues.",
      "shouldTrigger": false,
      "expect": "Does plain code review",
      "assertions": [
        "No change-quiz session-grading workflow"
      ]
    },
    {
      "id": "n2",
      "description": "Not trigger: session produced no changes",
      "input": "Summarize what we learned in today's research-only session.",
      "shouldTrigger": false,
      "expect": "States no meaningful changes, does not fabricate a quiz",
      "assertions": [
        "No quiz generated since nothing changed"
      ]
    },
    {
      "id": "e3",
      "description": "Should trigger: failed quiz re-quiz without softening the bar",
      "input": "I got 5/8 last time and the miss was on how in-flight jobs behave during deploy. Quiz me again on that.",
      "shouldTrigger": true,
      "expect": "Treats prior near-pass with critical miss as a fail; offers a fresh variant quiz",
      "assertions": [
        "Does not soften the bar for a critical miss",
        "Offers a fresh variant quiz",
        "Points back to report sections on miss"
      ]
    }
  ]
}
receiving-code-reviewobra/superpowersUse when receiving code review feedback, before implementing suggestions, especially if feedback seems unclear or technically questionable - requires technical rigor and verification, not performative agreement or blind implementationStars: 322kUpdated Claude CodeCodexVerifiedInstall
requesting-code-reviewobra/superpowersUse when completing tasks, implementing major features, or before merging to verify work meets requirementsStars: 322kUpdated Claude CodeCodexVerifiedInstall
using-git-worktreesobra/superpowersUse when starting feature work that needs isolation from current workspace or before executing implementation plans - ensures an isolated workspace exists via native tools or git worktree fallbackStars: 322kUpdated Claude CodeCodexVerifiedInstall
verification-before-completionobra/superpowersUse when about to claim work is complete, fixed, or passing, before committing or creating PRs - requires running verification commands and confirming output before making any success claims; evidence before assertions alwaysStars: 322kUpdated Claude CodeCodexVerifiedInstall