proof-checker

Rigorous mathematical proof verification and fixing workflow. Reads a LaTeX proof, identifies gaps via cross-model review (external reviewer backend, ultra reasoning), fixes each gap with full derivations, re-reviews, and generates an audit report. Use when user says "检查证明", "verify proof", "proof check", "审证明", "check this proof", or wants rigorous mathematical verification of a theory paper.

Quality

88%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Passed

No findings from the security scan

Quality

Content

77%Weight 40%Scale 1-5

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

A highly actionable, well-sequenced workflow with strong validation feedback loops, but it is monolithic and verbose: substantial material that belongs in separate reference files is inlined, and the opt-in discipline is repeated across multiple sections.

Suggestions

Move the issue taxonomy (20 categories), side-condition checklists, and the PROOF_AUDIT.json schema contracts into bundled reference files under references/, keeping SKILL.md as an overview that links to them one level deep.

Consolidate the deep-fix and restatement-check opt-in rules into a single authoritative section; the Phase addenda, Key Rules, and Submission Artifact Emission sections currently restate the same verdict-crosstalk and additivity guarantees multiple times.

Tighten the Constants and Reviewer Calling Convention blocks by removing the inline capability-fallback prose and pointing once to shared-references/reviewer-routing.md instead of re-explaining the fallback chain in the body.

Dimension	Reasoning	Score
Conciseness	The core protocol is high-signal, but the deep-fix and restatement-check opt-in rules are restated across three or more locations (Phase addenda, dedicated sections, Key Rules, and Submission Artifact Emission), which is padded repetition that could be consolidated.	3 / 5
Actionability	Provides copy-paste-ready guidance throughout: exact MCP calls with model/config parameters, the full reviewer prompt in a code block, concrete bash commands (pdflatex, research_wiki.py with full flags), and complete JSON output schemas.	5 / 5
Workflow Clarity	A clearly sequenced Phase 0–5.5 workflow with explicit validation checkpoints (compile check, acceptance gate, blind independent re-review, regression audit) and feedback loops ('If not met, repeat Phases 2-3', 're-enter Phase 2' on new blind-review issues).	5 / 5
Progressive Disclosure	Section structure and external shared-references links are clearly signaled, but the issue taxonomy, side-condition checklists, schema contracts, and opt-in mode specs are all inlined in a single ~860-line file with no bundle reference files to split them into.	3 / 5
	Total	16 / 20 Passed

Description

100%Weight 40%Scale 1-5

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

An exemplary description: third-person voice, concrete multi-step capabilities, comprehensive bilingual trigger terms, and an explicit 'Use when' clause. It clearly communicates both what the skill does and when to invoke it.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete actions — 'Reads a LaTeX proof, identifies gaps via cross-model review', 'fixes each gap with full derivations, re-reviews, and generates an audit report' — giving comprehensive coverage of what the skill does.	5 / 5
Completeness	Explicitly answers both 'what' (proof verification and fixing workflow with concrete steps) and 'when' ('Use when user says ...') with concrete trigger phrases.	5 / 5
Trigger Term Quality	Comprehensive natural trigger terms across both English and Chinese synonyms — 'verify proof', 'proof check', 'check this proof', '检查证明', '审证明' — plus the descriptive phrase 'rigorous mathematical verification of a theory paper'.	5 / 5
Distinctiveness Conflict Risk	Occupies a clear niche — rigorous mathematical proof verification via cross-model adversarial review — with highly specific triggers unlikely to collide with other skills.	5 / 5
	Total	20 / 20 Passed

Validation

75%

Warnings & errors only

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 12 / 16 Passed

Validation for skill structure

Criteria	Description	Result
skill_md_line_count	SKILL.md is long (867 lines); consider splitting into references/ and linking	Warning
allowed_tools_field	'allowed-tools' contains unusual tool name(s)	Warning
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning
relative_links	Relative link issues: 4 suspicious	Warning

	Total	12 / 16 Passed

Repository: wanshuiyin/Auto-claude-code-research-in-sleep
Path: skills/proof-checker/SKILL.md
Commit: 0472e53

Reviewed: 5 days ago

Table of Contents

Discovery Implementation Validation

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.