ClawHub
Back to skill

Security audit

**文件**：`wenyan-compress.skill` **大小**：~7KB --- **包含内容：** 1. **SKILL.md** — 主触发逻辑和两种工作模式 2. **references/compress-map.md** — 压缩映射表（动作/状态/情绪/主题词/时间戳） 3. **references/grammar.md** — 文言语法速查（句式/虚词/快速压缩公式） 4. **references/dict.md** — 典故成语对照表（诸子百家+近现代名人） --- **核心价值：** - 省Token也是个不错的赛道，中文语言，典故，成语，文言文都是压缩包，记忆压缩也可以参考。 **使用说明：** Skill 的核心设计： - 省Token也是个不错的赛道，中文语言，典故，成语，文言文都是压缩包，记忆压缩也可以参考。 - 触发：用户说"用文言文"、要求存档/压缩、或写入日志时自动激活 - 模式A：对外输出，追求文辞雅驯 - 模式B：内部存档，用 时间·主语·动作·状态 格式高效压缩 - 三大参考文件：映射表 + 语法速查 + 典故词典（诸子百家都有） 安装后直接说"用文言文回答"即可体验。 **使用示例：** 问题：用文言文帮我概括下十五五的规划重点 回答： 十五五规划（即第十四个五年规划收官与第十五个五年规划纲要，2026-2030）之要旨，试以文言述之： 十五五规划要义 当世之局，变乱交织，十五五之策，以创新驱动为核。科技自立，关乎存亡；人工智能、量子计算、深海深空，皆兵家必争之地，此新质生产力之谓也。 内循环为主，外开放为辅——内需畅则根基固，外联广则财源通，二者相济，不可偏废。 绿色转型，势在必行。碳达峰、碳中和，非一时之功，乃百年之计。清洁能源、储能技术、低碳制造，皆为急务。 乡村振兴，与城并举。县域经济、特色产业、数字农桑，皆可兴焉。 共同富裕，非均贫富，乃崇教育、平机会、缩差距，使民各安其业、各得其所。 国家安全，至重至要。粮食、能源、信息、生物，四域皆需有备。 五字诀： 创之以科技，循之以绿色，畅之以内需，富之以共惠，固之以国安。 欲详某一领域（如AI、芯片、新能源等）之专项规划，可再询之。

Security checks for vulnerabilities and agentic risk

Overview

This is a text-only Chinese classical-language compression skill with some broad activation guidance, but no hidden code, network access, credential handling, or destructive behavior.

Install only if you want Chinese classical-language compression for summaries, logs, or memory notes. Review when it activates, because broad requests like summaries or record-keeping may cause future notes to be stored in a terse Wenyan format.

A.I.G

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs

SkillSpector

Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (6)

Vague Triggers

High
Category
Not specified by scanner
Confidence
95% confidence
Finding

The trigger conditions are overly broad and include common phrases such as '帮我记录一下', '总结', long-conversation summarization, and any write to memory/log files. This can cause the skill to activate in ordinary contexts without clear user intent, unexpectedly rewriting content into compressed classical Chinese and potentially corrupting logs, summaries, or stored memory in ways that reduce transparency and auditability.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
85% confidence
Finding

The skill definition centers on converting arbitrary Chinese responses into classical Chinese and enabling reversible compressed archives, but it does not require a clear user language preference beyond trigger matching. In practice this can override expected output format, reduce readability, and introduce misunderstanding when users did not explicitly consent to language transformation.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

This markdown file defines mappings for phrases like “帮我做”, “帮我找”, “帮我写”, “帮我查”, and “帮我算”, which are generic requests commonly used in everyday conversation. Without any scope constraints, activation rules, or exclusion examples, these phrases could cause unintended invocation if used as skill triggers.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

This markdown file is presented entirely in Chinese and titled as an allusion dictionary, with no indication that language selection is optional or that a Chinese-only locale is intentionally required. Under the policy rule for natural-language violations, a skill that forces a specific language without user opt-in should be flagged.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
84% confidence
Finding

The entire file specifies a Chinese colloquial-to-classical compression scheme and presents it as the operative format, with no indication that users may opt into another language or locale. This can violate language-choice policy when a skill imposes a specific language format without user selection or justification.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
93% confidence
Finding

This markdown file is entirely a Chinese/Wenyanwen grammar cheatsheet and does not offer any language choice or note that the content is intended only for users who opted into Chinese output. Under the policy, forcing a specific language without user opt-in is a natural-language policy violation.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.