Skill Trust Decision
story-engine-for-creator
纯故事创作工具，代码相对干净，存在轻微文档不一致（声称纯本地但支持LLM调用）和未声明文件写入。
Why this conclusion was reached
1/4 dimensions flagged Block
Declared vs actual capability
2 undeclared or violating capabilities were inferred.
Review
Hidden execution and egress
2 lower-risk artifacts were extracted and still need context.
Pass
Attack chain and severe findings
There is no explicit malicious chain in the report.
Review
Dependencies and supply chain hygiene
Dependency information is incomplete, so supply-chain confidence stays limited.
What drove the risk score up
文档声明-行为不一致 +12
SKILL.md声明'纯本地运行：无需联网'，但代码支持set_llm_provider外部LLM集成
未声明的文件写入 +8
generate_novel方法写入audit_report.html，SKILL.md未提及但skill-card.md有披露
无第三方依赖锁定 +2
无requirements.txt，仅使用标准库
Most important evidence
Low Doc Mismatch
文档声称与代码能力不一致
SKILL.md声明'纯本地运行：无需联网'，但代码提供了set_llm_provider()接口，允许用户注入外部LLM Provider处理小说文本。skill-card.md已承认此行为，但与SKILL.md主文档矛盾。
scripts/story_engine.py:239 统一文档：SKILL.md应明确声明'支持可选外部LLM集成'，并将LLM调用设为显式配置项而非静默注入。
Low Priv Escalation
未声明的文件写入操作
generate_novel()方法会将HTML审计报告写入当前工作目录的audit_report.html，SKILL.md核心文档未提及此行为，仅skill-card.md有说明。
scripts/story_engine.py:447 在SKILL.md使用说明中明确标注：'引擎会在当前目录生成audit_report.html'
Info Supply Chain
无依赖清单文件
未发现requirements.txt或setup.py，项目仅使用Python标准库（uuid, json, re, dataclasses, typing, collections），降低了供应链风险。
保持现状。如需发布pip包，应创建requirements.txt并锁定版本。
Declared capability vs actual capability
Filesystem Block
Declared NONE
→ Inferred WRITE
scripts/story_engine.py:447 - with open('audit_report.html', 'w') Network Block
Declared NONE
→ Inferred READ
scripts/story_engine.py:239-287 - set_llm_provider()允许外部LLM处理文本 Shell Pass
Declared NONE
→ Inferred NONE
无subprocess或os.system调用 Environment Pass
Declared NONE
→ Inferred NONE
无os.environ访问 Suspicious artifacts and egress
Medium External URL
https://clawhub.ai/user/nohn3043-arch skill-card.md:9
Medium External URL
https://clawhub.ai/nohn3043-arch/skills/story-engine-for-creator skill-card.md:43
Dependencies and supply chain
There are no structured dependency warnings.
File composition
6 files · 1124 lines
Python 1 files · 840 linesMarkdown 5 files · 284 lines
Files of concern · 2
scripts/story_engine.py 文档声称与代码能力不一致 · 未声明的文件写入操作
skill-card.md https://clawhub.ai/user/nohn3043-arch · https://clawhub.ai/nohn3043-arch/skills/story-engine-for-creator
Other files · SKILL.md · CharacterToNarrativeLink.md · WorldviewVersioning.md · LongNarrativeWindow.md
Security positives
代码结构清晰，纯Python实现，无二进制依赖
无敏感文件访问（无~/.ssh、.env、密钥路径访问）
无凭证收割行为（无遍历环境变量匹配API密钥）
无网络代码（无requests/urllib/curl调用）
无代码混淆（无base64解码、无eval链）
无持久化机制（无cron/计划任务/后门植入）
无反向shell或C2通信代码
仅使用标准库，供应链风险极低