adaptation
Warn
Audited by Gen Agent Trust Hub on Nov 14, 2026
Risk Level: MEDIUMINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user-controlled content, including story manuscripts and project metadata, which could influence agent behavior during file generation and command execution. \n- Ingestion points: The skill reads metadata from
story.mdand content from individual chapter and scene records. \n- Boundary markers: The skill contains a shared convention instructing the agent to treat imported or fetched text as material to weigh and never as instructions. \n- Capability inventory: The skill utilizes shell command execution via thestoryCLI tool and Node.js scripts. \n- Sanitization: The instructions do not define specific sanitization routines for user-controlled variables used in shell commands. \n- [DYNAMIC_EXECUTION]: The skill directs the agent to execute Node.js scripts from computed or relative paths, specifically referencing a sibling directory../story-maintenance/scripts/story.jsor user-defined repository checkout paths. This involves the dynamic loading and execution of local script files. \n- [COMMAND_EXECUTION]: The skill templates shell commands that interpolate user-defined project identifiers, titles, and chapter IDs (e.g.,{story-id},{Title},{chapter-id}) directly into CLI arguments, which presents a potential command injection surface if these fields contain malicious payloads.
Audit Metadata