aiskillstore/marketplace / pending/dontbesilent2025/dbs-ai-check
dbs-ai-check
Verified扫描文案中的 AI 写作特征并输出检测报告,默认只诊断不改写;用户看完报告说「我想改」后,逐条追问写法背后的意图并给出修改方向。用户要求检查 AI 味、AI 痕迹或机器化表达时使用。
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 12.5 KB |
| agents/openai.yaml | config | 278 B |
| skill-report.json | config | 15.2 KB |
skill-report.json
View on GitHub{
"schema_version": "2.0",
"meta": {
"generated_at": "2026-11-03T15:44:28.275Z",
"slug": "dontbesilent2025-dbs-ai-check",
"source_url": "https://github.com/dontbesilent2025/dbskill/tree/eacbecfa6b10dd3475da0a1edb84c3b8b3084d3c/skills/dbs-ai-check",
"source_ref": "eacbecfa6b10dd3475da0a1edb84c3b8b3084d3c",
"model": "codex",
"analysis_version": "3.0.0",
"source_type": "community",
"content_hash": "d23932e3bb42d8e9301d4ecd6ae572f674d189d86c712dc325f2ba2d5976f92b",
"tree_hash": "01d7c0c2a708add499dff65e25f6f950ce6d3404e064cce5a3cd65ab920b9814",
"provenance": {
"model_effective": "codex",
"fallback_chain": [
{
"agent": "codex",
"outcome": "success"
}
],
"analysis_version": "3.0.0"
},
"upstream_version_raw": null,
"upstream_version_normalized": null,
"upstream_version_source": "none",
"upstream_version_status": "missing",
"upstream_commit_sha": "eacbecfa6b10dd3475da0a1edb84c3b8b3084d3c"
},
"skill": {
"name": "dbs-ai-check",
"description": "扫描文案中的 AI 写作特征并输出检测报告,默认只诊断不改写;用户看完报告说「我想改」后,逐条追问写法背后的意图并给出修改方向。用户要求检查 AI 味、AI 痕迹或机器化表达时使用。",
"summary": "Review writing against 22 AI-style traits, then guide revisions through questions about the author's intent.",
"icon": "📦",
"author": "dontbesilent2025",
"category": "writing",
"tags": [
"writing-review",
"ai-writing",
"chinese-copywriting",
"editorial-feedback",
"author-voice"
],
"supported_tools": [
"claude",
"codex",
"claude-code"
],
"risk_factors": [
"external_commands"
]
},
"security_audit": {
"risk_level": "safe",
"is_blocked": false,
"safe_to_publish": true,
"analysis_status": "ok",
"summary": "All seven static findings are false positives: backticks mark Markdown examples or chat commands, and the entropy warning concerns readable Chinese text. No evidence found of shell execution, concealed payloads, data exfiltration, or security-review manipulation in the reviewed files.",
"remediation": [],
"risk_factor_evidence": [
{
"factor": "external_commands",
"evidence": [
{
"file": "SKILL.md",
"line_start": 46,
"line_end": 58
},
{
"file": "SKILL.md",
"line_start": 58,
"line_end": 65
},
{
"file": "SKILL.md",
"line_start": 65,
"line_end": 74
},
{
"file": "SKILL.md",
"line_start": 74,
"line_end": 246
},
{
"file": "SKILL.md",
"line_start": 246,
"line_end": 268
},
{
"file": "SKILL.md",
"line_start": 268,
"line_end": 268
}
]
}
],
"static_findings": [
{
"id": "external_commands:SKILL.md:46:ruby-shell-backtick-execution",
"category": "external_commands",
"severity": "medium",
"pattern": "Ruby/shell backtick execution",
"file": "SKILL.md",
"line_start": 46,
"line_end": 58,
"snippet": "```"
},
{
"id": "external_commands:SKILL.md:58:ruby-shell-backtick-execution",
"category": "external_commands",
"severity": "medium",
"pattern": "Ruby/shell backtick execution",
"file": "SKILL.md",
"line_start": 58,
"line_end": 65,
"snippet": "`特征 #N 特征名 严重度`"
},
{
"id": "external_commands:SKILL.md:65:ruby-shell-backtick-execution",
"category": "external_commands",
"severity": "medium",
"pattern": "Ruby/shell backtick execution",
"file": "SKILL.md",
"line_start": 65,
"line_end": 74,
"snippet": "`特征 #N 特征名 严重度`"
},
{
"id": "external_commands:SKILL.md:74:ruby-shell-backtick-execution",
"category": "external_commands",
"severity": "medium",
"pattern": "Ruby/shell backtick execution",
"file": "SKILL.md",
"line_start": 74,
"line_end": 246,
"snippet": "```"
},
{
"id": "external_commands:SKILL.md:246:ruby-shell-backtick-execution",
"category": "external_commands",
"severity": "medium",
"pattern": "Ruby/shell backtick execution",
"file": "SKILL.md",
"line_start": 246,
"line_end": 268,
"snippet": "- 用户的文案本身选题有问题 → 明确指出选题信号,把它写进本轮结论。检测结束后,用户想继续时输入 `/dbs`,由主路由判断下一步。"
},
{
"id": "external_commands:SKILL.md:268:ruby-shell-backtick-execution",
"category": "external_commands",
"severity": "medium",
"pattern": "Ruby/shell backtick execution",
"file": "SKILL.md",
"line_start": 268,
"line_end": 268,
"snippet": "完成当前任务后直接结束。用户的请求超出本 Skill 的范围、任务完成后用户的材料里出现明确的后续需求,或用户询问下一步时,用一句话说明还需要哪种能力,不点名其他 Skill;当前环境已经安装 `/d"
},
{
"id": "obfuscation:SKILL.md:1:heuristic-high-file-entropy-7-77-bits-possible-b",
"category": "obfuscation",
"severity": "high",
"pattern": "[HEURISTIC] High file entropy (7.77 bits) - possible binary/encrypted content",
"file": "SKILL.md",
"line_start": 1,
"line_end": 1,
"snippet": "File: SKILL.md"
}
],
"finding_verdicts": [
{
"id": "external_commands:SKILL.md:46:ruby-shell-backtick-execution",
"verdict": "false_positive",
"confidence": 1,
"reason": "Line 46 opens a Markdown fence containing a sample editorial report. The enclosed text contains report placeholders, not Ruby or shell instructions."
},
{
"id": "external_commands:SKILL.md:58:ruby-shell-backtick-execution",
"verdict": "false_positive",
"confidence": 1,
"reason": "Line 58 uses inline backticks for the report's trait number, name, and severity label. This is display formatting inside an example, not command substitution."
},
{
"id": "external_commands:SKILL.md:65:ruby-shell-backtick-execution",
"verdict": "false_positive",
"confidence": 1,
"reason": "Line 65 repeats the inline Markdown label for the second sample report entry. No execution mechanism or executable command appears."
},
{
"id": "external_commands:SKILL.md:74:ruby-shell-backtick-execution",
"verdict": "false_positive",
"confidence": 1,
"reason": "Line 74 closes the Markdown fence opened at line 46. The following sections describe editorial rules and questions, not shell execution."
},
{
"id": "external_commands:SKILL.md:246:ruby-shell-backtick-execution",
"verdict": "false_positive",
"confidence": 0.99,
"reason": "Line 246 presents /dbs in Markdown backticks as a user-entered chat command for optional follow-up. It does not instruct terminal execution."
},
{
"id": "external_commands:SKILL.md:268:ruby-shell-backtick-execution",
"verdict": "false_positive",
"confidence": 0.99,
"reason": "Line 268 conditionally suggests the installed /dbs chat router when follow-up is relevant. The backticks format its name, without invoking a shell or installing tools."
},
{
"id": "obfuscation:SKILL.md:1:heuristic-high-file-entropy-7-77-bits-possible-b",
"verdict": "false_positive",
"confidence": 0.99,
"reason": "SKILL.md contains readable Chinese prose, YAML front matter, and Markdown tables and examples throughout. No encoded or encrypted payload is present; the entropy heuristic misclassifies multilingual text."
}
],
"semantic_findings": [],
"critical_findings": [],
"high_findings": [],
"medium_findings": [],
"low_findings": [],
"dangerous_patterns": [],
"files_scanned": 2,
"total_lines": 274,
"audit_model": "codex",
"audited_at": "2026-11-03T15:44:28.275Z",
"agent_auto_install_policy": "allowed",
"manual_install_policy": "allowed"
},
"content": {
"user_title": "Review AI-Style Writing and Refine Your Voice",
"value_statement": "Formulaic writing can obscure an author's intent and weaken editorial feedback. This skill identifies specific style patterns and guides revisions through questions before changing text.",
"seo_keywords": [
"Claude",
"Codex",
"Claude Code",
"AI writing review",
"Chinese copywriting",
"editorial feedback",
"author voice",
"writing style analysis",
"genre-aware editing",
"guided revision"
],
"actual_capabilities": [
"Review submitted writing against 22 defined AI-style traits.",
"Quote each flagged passage in text order and identify its trait number and signal strength.",
"Adjust criteria for short video scripts, long articles, social posts, and formal writing.",
"Report when no clear AI-style patterns are found instead of forcing findings.",
"Ask one intent-focused question at a time, prioritizing stronger signals after the user requests revision guidance.",
"Suggest specific revisions after the author answers and explicitly requests changes."
],
"limitations": [
"Style signals do not prove AI authorship or provide validated detection accuracy.",
"The checklist emphasizes Chinese writing, although responses can follow the user's language.",
"The default workflow diagnoses writing and does not immediately rewrite it.",
"The skill does not provide comprehensive topic evaluation, fact checking, or plagiarism detection."
],
"use_cases": [
{
"title": "Review a social media draft",
"description": "Identify repetitive rhetoric and formulaic transitions while accounting for the conventions of short social posts.",
"target_user": "Social media writers"
},
{
"title": "Prepare passage-level editorial feedback",
"description": "Review a long article with quoted passages, trait labels, and signal strengths before discussing changes with the author.",
"target_user": "Editors"
},
{
"title": "Preserve intent during revision",
"description": "Work through flagged passages one question at a time to clarify the intended meaning before requesting specific changes.",
"target_user": "Independent authors"
}
],
"prompt_templates": [
{
"title": "Check a draft",
"prompt": "Review this draft for AI-style writing patterns. Quote relevant passages and explain each concern without rewriting: [draft]",
"scenario": "Beginner: obtain a first diagnostic report."
},
{
"title": "Apply genre-specific criteria",
"prompt": "Review this short video script using the appropriate genre exceptions. Identify specific patterns without treating every memorable closing sentence as suspicious: [script]",
"scenario": "Intermediate: reduce genre-related false positives."
},
{
"title": "Explore revision intent",
"prompt": "I want revision guidance for the report below. Start with the strongest signal, ask one question about my intent, and wait: [report]",
"scenario": "Intermediate: begin an interactive editorial discussion."
},
{
"title": "Revise after clarification",
"prompt": "Using my answers below, revise only the flagged passage. Preserve my intended meaning and stated preferences: [passage, report, and answers]",
"scenario": "Advanced: request a focused revision after clarifying intent."
}
],
"output_examples": [
{
"input": "Review a long article containing repeated sentences beginning with 'However', 'In fact', and 'It is worth noting'.",
"output": [
"Genre: long article.",
"Quoted openings: 'However', 'In fact', and 'It is worth noting'.",
"Trait #17: repeated connectors in fixed positions. Medium signal.",
"These repeated openings make the transitions formulaic. Review their frequency in the full article before deciding which ones are unnecessary."
]
},
{
"input": "I want revision guidance after a report flags several equally long parallel sentences.",
"output": [
"Start with trait #3: uniform parallel phrasing.",
"Which sentence contains the point you most want the reader to remember?",
"Revision guidance pauses here until the author answers."
]
}
],
"best_practices": [
"Provide the complete passage and identify its genre before requesting a review.",
"Treat each finding as editorial feedback and verify it against the passage's purpose and context.",
"Answer the intent questions before requesting specific revisions."
],
"anti_patterns": [
"Use style findings as proof that a person used AI.",
"Request rewriting designed to disguise AI authorship.",
"Apply every checklist item mechanically without considering genre or normal rhetorical usage."
],
"faq": [
{
"question": "Does this skill prove that a text was generated by AI?",
"answer": "No. It identifies stylistic signals that human writers can also produce, so findings do not establish authorship."
},
{
"question": "Will it rewrite my draft immediately?",
"answer": "No. It reports issues first, asks about your intent when requested, and provides specific changes only after clarification and explicit permission."
},
{
"question": "What does the diagnostic report include?",
"answer": "It quotes flagged passages in text order and explains each concern with a trait number and signal strength."
},
{
"question": "Which writing genres does it consider?",
"answer": "It defines adjustments for short video scripts, long articles, social posts, and academic or formal writing."
},
{
"question": "Can I use it with English text?",
"answer": "It instructs the assistant to answer English requests in English, but many checklist examples and thresholds are specific to Chinese writing."
},
{
"question": "What happens when no clear patterns are found?",
"answer": "It should state that no clear issues were found rather than inventing problems to fill a report."
}
]
},
"file_structure": [
{
"name": "agents",
"type": "dir",
"path": "agents",
"children": [
{
"name": "openai.yaml",
"type": "file",
"path": "agents/openai.yaml",
"lines": 5
}
]
},
{
"name": "SKILL.md",
"type": "file",
"path": "SKILL.md",
"lines": 269
}
]
}
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| slack-gif-creatoranthropics/skillsKnowledge and utilities for creating animated GIFs optimized for Slack. Provides constraints, validation tools, and animation concepts. Use when users request animated GIFs for Slack like "make me a GIF of X doing Y for Slack." | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| ponytailDietrichGebert/ponytailLazy senior dev mode for any coding task: the smallest change that fully solves it, and a reply a busy human understands in one read. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailHarvest every shortcut: comment into one debt ledger, so deferrals get tracked instead of forgotten. One-shot report. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailList every `shortcut:` comment (and older `ponytail:` ones) as a debt ledger. One-shot report, changes nothing. Use for "ponytail debt", "what did ponytail defer", "list the shortcuts", /ponytail-debt. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |