story
Pass
Audited by Gen Agent Trust Hub on Oct 11, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill fetches update information, release tags, and release notes from the official GitHub repository for version comparison and notification purposes.
- [COMMAND_EXECUTION]: The skill launches a local Node.js server to provide a project dashboard and executes a Python script to manage structured author preferences and story metadata.
- [INDIRECT_PROMPT_INJECTION]:
- Ingestion points: The skill spawns agents (e.g.,
story-explorer) that read and process content from local project files, including story settings, character descriptions, and plot tracking documents (设定/,追踪/). - Boundary markers: The skill does not explicitly define delimiters or boundary markers for the data ingested by sub-agents, relying on the agents' internal instructions for safety.
- Capability inventory: The skill can spawn sub-agents, perform file read/write operations via a local HTTP server, and execute shell commands for version checking and dashboard management.
- Sanitization: The dashboard server incorporates path validation to prevent traversal attacks and restricts connections to local loopback addresses by default to prevent unauthorized network access.
Audit Metadata