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

90%

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

81%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.

The body is a highly actionable, rigorously validated multi-phase workflow with excellent feedback loops and concrete code/prompts throughout. Its main weakness is conciseness: opt-in-flag semantics are restated across four sections each, and large reference-grade content (taxonomy, schemas, algorithms) is inlined rather than pushed into the shared-references files it already links to.

Suggestions

Collapse the deep-fix opt-in semantics (currently restated in the Phase 1 addendum, the Deep-Fix Mode section, Key Rules, and Submission Artifact Emission) into a single canonical block referenced once, keeping only a one-line pointer elsewhere.

Do the same for the restatement-check semantics (Phase 3.6, Key Rules, Output Files, Submission Artifact Emission) — define once, point to it from the other three locations.

Move the 20-category issue taxonomy, the side-condition checklist table, and the full assurance-contract JSON schema into shared-references files (e.g. issue-taxonomy.md, assurance-contract.md already linked) and keep only a summary plus link in SKILL.md to reduce the inlined bulk.

Dimension	Reasoning	Score
Conciseness	The body is dense with substantive domain instructions rather than common-knowledge fluff, but the deep-fix and restatement-check opt-in semantics are each restated verbatim across four separate sections (Phase 1 addendum, dedicated section, Key Rules, Submission Artifact Emission), which is unnecessary repetition that could be tightened; this lands at anchor 3 rather than 2 because each instance carries real content rather than padding.	3 / 5
Actionability	Fully executable, copy-paste-ready guidance throughout — exact reviewer prompts in code blocks, concrete MCP calls with pinned model/config JSON, runnable bash (`pdflatex ... \| grep -E "Error\|Warning\|undefined"`, `python3 "$WIKI_SCRIPT" add_claim ...`), and complete output JSON schemas, matching anchor 5's bar for common cases.	5 / 5
Workflow Clarity	A batch/destructive skill with exemplary validation: objective acceptance gate, per-fix compile check, Phase 3 re-review convergence loop, Phase 3.5 global closure + blind independent re-review for FATAL/CRITICAL, regression audit (DAG acyclicity, counterexample re-run, assumption-delta), and an explicit unrecoverable protocol — satisfying the destructive-skill validation requirement and matching anchor 5's explicit-checkpoint/checklist descriptor.	5 / 5
Progressive Disclosure	Good section structure (numbered phases, clear ## headings) and references that are clearly signaled via markdown links to shared-references/*.md files; however substantial reference-grade content (the 20-category issue taxonomy, full assurance-contract JSON schema, the entire deep-fix and Phase 3.6 algorithms) is inlined rather than split into the referenced files, which is more than the minor gaps of anchor 5 yet better-organized than anchor 3.	4 / 5
	Total	17 / 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.

The description is exemplary: it pairs concrete workflow actions with bilingual, synonym-rich natural trigger phrases, explicitly stating both what the skill does and when to use it within a clearly distinct niche. It hits the top anchor on every dimension with no over-claims or vague fluff.

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 the verification workflow rather than the minimal 1-2 actions at anchor 3.	5 / 5
Completeness	Explicitly answers both "what" (verification + fixing workflow with listed actions) and "when" ("Use when user says...") with concrete trigger phrases, matching the anchor-5 example structure exactly; not the weak/implied "when" of anchor 4.	5 / 5
Trigger Term Quality	Comprehensive bilingual natural trigger phrases — "检查证明", "verify proof", "proof check", "审证明", "check this proof" — covering synonyms (verify/check, proof/proof check) that users would actually say, matching anchor 5's comprehensive-synonyms bar.	5 / 5
Distinctiveness Conflict Risk	Occupies a clear niche — rigorous mathematical proof verification — with distinctive triggers ("verify proof", "proof check") unlikely to fire for unrelated skills, matching anchor 5's clear-niche/minimal-conflict descriptor.	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: f4f20f9

Reviewed: 1 day 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.