release-illustration-planner

Pass

Audited by Gen Agent Trust Hub on Nov 4, 2026

Risk Level: SAFEDYNAMIC_EXECUTIONCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONNO_CODE
Full Analysis
  • [DYNAMIC_EXECUTION]: The Python script scripts/planner.py dynamically modifies the system path using sys.path.insert(0, ...) to import local modules from computed paths. These paths include sibling directories such as handdraw-style-prompter and handraw-fabu, creating a dependency on a specific local project structure and environment variables like RELEASE_SCRIPTS_DIR.
  • [COMMAND_EXECUTION]: The skill performs extensive file system operations through its Python script. It creates directories using Path.mkdir(parents=True, exist_ok=True) and writes structured data and documentation files (cases_manifest.json, cases_manifest.md) to the local disk based on user-supplied parameters like --article-title.
  • [INDIRECT_PROMPT_INJECTION]: The planner script implements an ingestion pipeline that reads content from markdown files in the business/ directory. It uses regular expressions to extract domain information, materials, and reasons, which are then interpolated directly into the final prompts sent to the image generation engine. This represents a vulnerability surface where maliciously crafted local files could influence the resulting image generation instructions.
  • [DATA_EXPOSURE]: The skill accesses several internal configuration files, including styles.json, layouts.json, and colors.json, located within the local project environment to perform its planning logic and ensure compliance with project-specific art standards.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 4, 2026, 03:51 PM