story-long-write

Warn

Audited by Gen Agent Trust Hub on Jul 26, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to execute bundled Node.js scripts including scripts/check-ai-patterns.js, scripts/normalize-punctuation.js, and scripts/check-degeneration.js to validate and modify generated story files. Additionally, it utilizes Python shell commands to calculate precise character counts during the validation phase.
  • [EXTERNAL_DOWNLOADS]: The skill references an external repository (https://github.com/worldwonderer/oh-story-claudecode) in its metadata as its source. It also involves managing and importing story assets from directories such as 拆文库/, which serves as a data source for competitive analysis and benchmarking.
  • [PROMPT_INJECTION]: The skill ingests untrusted data from external files such as 选题决策.md and reference assets in the 拆文库/ directory. Due to the lack of explicit boundary markers for this external content and the agent's significant capabilities (file writing and script execution), there is a risk of indirect prompt injection where instructions embedded in these processed files could influence agent behavior.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Jul 26, 2026, 10:44 PM
Security Audit — agent-trust-hub — story-long-write