lukehsiao/dotfiles/hegel-review
Claude Code and agent skill by lukehsiao/dotfiles
Trust tier 3
- community
Source check pending
Checking that the SKILL.md file is still in the repository.
Review property-based tests for the failure modes that make them weak or misleading. Use after writing property-based tests with hegel (or proptest, quickcheck, fast-check, jqwik, etc.), when asked to review property-based tests, or before shipping a change to a PBT suite.
Install
npx vskill@latest install lukehsiao/dotfiles/hegel-reviewSKILL.md
Reviewing property-based tests
Check every test against this list. When a check fails, quote the offending code and give the smallest change that fixes it.
- Narrowed generators: drawn domains trimmed to dodge failures — ranges shrunk, characters filtered, sizes capped in the domain rather than at materialization, or a whole value class (subnormals, tiny magnitudes, exact bounds) excluded after it failed there, with or without an explanatory comment. The domain should be as wide as the contract.
- Fixed inputs in property clothing: a "property" that only ever sees one value or a few hand-picked cases.
- Missing direction: a parser or validator tested only on valid input; or invalid input checked only for "doesn't crash" when rejection should be asserted.
- More than one property per test: assertions about unrelated contracts sharing one test and one failure signal.
- Evidence-free properties: asserting behavior nothing documents or implies, or restating the implementation instead of a contract.
- Tolerance hacks: an epsilon widened until the test passes; approximate equality where the contract is exact. Equality, ordering, and hashing laws, exact roundtrips, and endpoint or identity claims (ratio 0 or 1 returning an endpoint) are exact contracts; a failed check on one dismissed as "a precision limitation", or a tolerance added after a failure was observed with no ledger note grounding it in the contract, is this hack applied during triage.
- Weakened oracles: asserting only
is_ok(), only a length, or only "doesn't panic" where the actual value could be checked cheaply. - `should_panic` or ignored tests: tests that pass while the bug is present, or are excluded from the default run.
- Config parameters fixed: constructor and configuration knobs pinned to defaults instead of generated.
- Untested variants: one property instantiated for one type or variant when the contract covers several; or the same property copy-pasted with drift.
- Silent-triage residue: commented-out asserts, deleted failing tests, TODOs hiding a red test.
- Resource bounds in the wrong place: recursion depth or collection size limited in the drawn domain instead of at materialization.
Security
- Security scanPassed, 90 of 100Pattern checks on the skill's files.
- Model reviewPassed
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingdot_claude/skills/hegel-review/SKILL.md
Popularity
- Repository stars
- 8
- Repository forks
- 2
- 7-day trend score
- 0
Stars and forks are for the whole GitHub repository, not this skill alone. Not refreshed yet.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/lukehsiao/dotfiles/hegel-review)