auto-paper-improvement-loop

Autonomously improve a generated paper via GPT-6-Astra xhigh review → implement fixes → recompile, for 2 rounds. Use when user says "改论文", "improve paper", "论文润色循环", "auto improve", or wants to iteratively polish a generated paper.

Quality

80%

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, dragged down by a duplicated reviewer prompt and a large inlined whitelist spec that should be offloaded to a reference file.

Suggestions

Factor the ~100-line edit-whitelist section (schema, resolution rules, glob semantics, forbidden-operation detectors) into a references/edit-whitelist.md file and keep only a brief summary plus a pointer in SKILL.md.

De-duplicate the Step 2 and Step 5 reviewer prompts: define the shared review template once and reference it, noting only the Round-2 'fresh, zero-context' delta inline.

Consolidate the REVIEWER_BIAS_GUARD explanation so it appears fully once (e.g. the Reviewer Independence Protocol) and is only briefly cross-referenced from Constants and Key Rules.

Dimension	Reasoning	Score
Conciseness	Mostly high-signal with no basic-concept padding, but the ~50-line reviewer prompt is duplicated nearly verbatim in Steps 2 and 5, and the REVIEWER_BIAS_GUARD rationale is restated across Constants, the Reviewer Independence Protocol, and Key Rules — tightening these would improve token efficiency.	3 / 5
Actionability	Provides copy-paste-ready bash (latexmk, pdfinfo, grep), an executable python normalization snippet, full spawn_agent prompts with model/effort, a YAML whitelist schema, and regex detector tables — fully executable and covers the common cases.	5 / 5
Workflow Clarity	Steps 0–9 are explicitly sequenced with validation checkpoints (recompile + verify 0 errors, restatement regression test, location-aware format check) and clear stop criteria/feedback loops for the destructive recompile-and-edit operations.	5 / 5
Progressive Disclosure	Well-sectioned but monolithic: no bundle files exist (references/scripts/assets absent) and the ~100-line edit-whitelist spec with detector tables is inlined content that would be better placed in a separate reference file.	3 / 5
	Total	16 / 20 Passed

Description

83%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.

A strong description: third-person, concrete, and explicit about both capability and trigger conditions with multilingual natural phrases. Minor room to surface secondary features and a few more trigger synonyms.

Dimension	Reasoning	Score
Specificity	Names the domain (paper improvement) and several concrete actions — 'review → implement fixes → recompile, for 2 rounds' — but does not enumerate secondary features like the edit-whitelist or kill-argument phase, leaving minor coverage gaps.	4 / 5
Completeness	Explicitly answers both 'what' (improve a paper via review→fix→recompile for 2 rounds) and 'when' (concrete 'Use when user says …' trigger phrases), matching the top anchor.	5 / 5
Trigger Term Quality	Provides good natural-trigger coverage in two languages ('改论文', 'improve paper', '论文润色循环', 'auto improve', 'iteratively polish'), though a few plausible phrasings like 'refine paper' are absent.	4 / 5
Distinctiveness Conflict Risk	The specific GPT-6-Astra/xhigh review→fix→recompile loop is a clear niche with minimal conflict risk, though 'improve paper' has minor overlap with adjacent paper skills like /auto-review-loop.	4 / 5
	Total	17 / 20 Passed

Validation

81%

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 — 13 / 16 Passed

Validation for skill structure

Criteria	Description	Result
skill_md_line_count	SKILL.md is long (619 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

	Total	13 / 16 Passed

Repository: wanshuiyin/Auto-claude-code-research-in-sleep
Path: skills/skills-codex/auto-paper-improvement-loop/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.