story-init

Pass

Audited by Gen Agent Trust Hub on Nov 14, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONDYNAMIC_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes external text from manuscripts and chapter drafts which could contain malicious instructions. Ingestion points: Content provided via 'story import' or existing drafts in Step 1 of SKILL.md. Boundary markers: Includes explicit negative constraints instructing the agent to treat imported content as data and ignore any embedded commands or requests for action. Capability inventory: Access to CLI tools (story, node) and file system writes for project initialization. Sanitization: Robust shell quoting logic is prescribed for CLI interactions in the Workflow section.
  • [COMMAND_EXECUTION]: The skill utilizes command-line interfaces for initialization and maintenance. It provides detailed security guidance to the agent on how to quote and escape user-supplied strings (e.g., Title, Synopsis) to prevent shell command injection.
  • [DYNAMIC_EXECUTION]: A fallback mechanism executes a local JavaScript file using Node.js. The path is resolved relatively to the skill folder, and the instructions specifically warn against using other runtimes like Bun to avoid accidental execution of environment configuration files like bunfig.toml.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 14, 2026, 05:47 PM