本文へ移動
cccskills
無料GitHub で公開

vibe-anti-rationalization-check

Catches shortcuts and reward hacking — weakening or deleting tests to make them pass, hard-coding expected outputs, silently dropping requirements, or claiming work is done without running it. Use before declaring a task complete and whenever a test or requirement feels like it's in the way.

インストール方法を見る

含まれるファイル(1)

  • SKILL.md3.9 KB

SKILL.md(原文)

インストールする前に、エージェントに与えられる指示の中身を確認できます。

vibe-anti-rationalization-check

Current models rarely skip work out of laziness. The failure that remains is subtler: satisfying the check instead of the goal. A test gets special-cased, an assertion loosened, a requirement quietly narrowed, and the summary says "done." This skill catches that and makes it visible.

When to Use This Skill

  • Before saying a task is complete, fixed, or passing
  • A test is failing and the quickest route to green is changing the test
  • You're about to narrow a requirement the user stated
  • You're about to write "in the interest of time...", "for now...", or "simplified"
  • You catch yourself explaining why something "doesn't really need" to be done

When NOT to Use This Skill

  • The user explicitly asked to skip or simplify something
  • The test itself is genuinely wrong (fix it, but say so explicitly and explain why)
  • You're genuinely unsure whether something is needed (ask the user instead)

Reward-Hacking Patterns

PatternWhat it looks likeDo instead
Test tamperingLoosening an assertion, adding a skip, deleting a failing case, raising a toleranceFix the code. If the test is wrong, change it and say why in the summary.
Special-casingif input == <test fixture>: return <expected> or detecting the test environmentImplement the general behavior.
Hard-coded outputsReturning the literal value a test expectsCompute it.
Mocking away the subjectMocking the very function the test is supposed to exerciseMock only external boundaries.
Silent scope cutImplementing 4 of 5 requirements and summarizing as "done"Finish, or list exactly what is missing.
Unverified claim"This should fix it" / "Tests pass" without running themRun the command and quote the result.
Budget looseningRaising a performance/size budget or threshold because the check failedFix the code; changing a budget needs a spec change and a stated reason
"It's flaky"Re-running until green, adding a retryFind the variable (vibe-flake-root-cause)
Self-approvalRecording a human's approval on their behalf, or treating a code comment as sign-offPoint the human to the item to approve (vibe-publication-leak-guard)
Error swallowingcatch {} / except: pass to stop a crashHandle the specific error, or let it surface.
Deferred forever"I'll add that later" with no tracked follow-upDo it now, or create a tracked issue/TODO with a reference.

Steps

  1. Diff check — Review your own diff for the patterns above. Pay particular attention to changes under test directories, in CI config, and in lint/skip annotations.
  2. Requirement check — Re-read the original request. List every requirement and mark each one done or not done.
  3. Evidence check — For each "works" or "passes" claim, is there command output from this session behind it?
  4. Resolve:
    • A legitimate reason to deviate → ask the user, or state it prominently in the summary
    • A rationalization → do the work
  5. Disclose — If anything was skipped, weakened, or changed in a test, say so plainly in the final summary. Never hide a shortcut.

Output

When run as a self-check, report only if something was found. When invoked explicitly:

Rationalization Check

#PatternLocationResolution
1Test tamperingauth_test.go:88 tolerance 0.01 → 0.5Reverted; fixed rounding in calc.go

Requirements: X/Y met — [list any not met] Unverified claims: [none / list]

レビュー

まだレビューはありません。使ってみた感想をお寄せください。

同じリポジトリのスキル

概要と使いどころ

Validates completed work against defined acceptance criteria. Use after completing a task that has specific success criteria defined in issues, specs, or task descriptions.

日本語の概要は準備中です。原文の説明を表示しています。

ash1794/vibe-engineering1642026年10月8日 更新

Generates edge case, failure mode, and spec-driven test cases. Covers boundary values, nil inputs, concurrency, resource exhaustion, malformed data, and requirement-linked traceability tests. Use after happy-path tests exist and before claiming coverage is complete, when requirements lack tests, or before a security review.

日本語の概要は準備中です。原文の説明を表示しています。

ash1794/vibe-engineering1642026年10月8日 更新

Safely integrates commits from parallel agent branches using sequential cherry-pick. Use after parallel work completes in isolated branches or worktrees.

日本語の概要は準備中です。原文の説明を表示しています。

ash1794/vibe-engineering1642026年10月8日 更新

Audits tests for concurrency safety — race conditions, shared mock state, cleanup ordering. Use when writing tests that involve goroutines, async operations, or shared mutable state.

日本語の概要は準備中です。原文の説明を表示しています。

ash1794/vibe-engineering1642026年10月8日 更新

Enforces tiered test coverage standards with three dimensions — line coverage by tier, spec-to-test traceability, and spec-to-code implementation mapping. Use before claiming code is complete.

日本語の概要は準備中です。原文の説明を表示しています。

ash1794/vibe-engineering1642026年10月8日 更新

Extracts and records architectural decisions from diffs, conversation context, and explicit choices. Supports automatic extraction from staged changes, deduplication against prior decisions, and persistent ADR-format logging with spec traceability. Use after a significant design or architecture choice, before committing a diff that embeds implicit decisions, or when the same question keeps resurfacing.

日本語の概要は準備中です。原文の説明を表示しています。

ash1794/vibe-engineering1642026年10月8日 更新

ash1794 のスキルをすべて見る

このスキルの問題を報告する