data-context-extractor
Pass
Audited by Gen Agent Trust Hub on Sep 25, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill acts as a generator that ingests database schemas and existing documentation files to produce new agent skills. This ingestion process is susceptible to indirect prompt injection if the source data contains malicious instructions designed to influence the generator's logic.\n
- Ingestion points: Metadata discovery via database schema tools and the manual loading of existing skill directories or zip files in Iteration Mode.\n
- Boundary markers: The instructions lack explicit delimitation or instructions for the agent to ignore natural language commands found within the processed metadata or files.\n
- Capability inventory: The skill is capable of writing files to the local system to build the skill structure and executing a local script (
scripts/package_data_skill.py) to zip the result.\n - Sanitization: While the utility script performs structural validation (checking for frontmatter and placeholders), it does not sanitize the textual content of the generated files for potentially malicious instructions.\n- [COMMAND_EXECUTION]: The skill workflows include a step to execute the provided
scripts/package_data_skill.pyutility. This script is used to package the generated documentation into a zip file for delivery to the user.
Audit Metadata