skill-maker
Pass
Audited by Gen Agent Trust Hub on Oct 16, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill generates new agent instructions from user-provided interview data. 1. Ingestion points: User responses and task notes are used to construct the SKILL.md body. 2. Boundary markers: No delimiters are added to protect sub-agents from user-injected text. 3. Capability inventory: The skill writes files and uses the Agent tool to execute generated instructions. 4. Sanitization: Only structural checks are performed.
- [COMMAND_EXECUTION]: The skill executes local Python scripts to automate skill authoring. Evidence: instructions run scripts/skill_maker.py and reference utility scripts in the plugin lib directory.
- [DYNAMIC_EXECUTION]: The skill dynamically generates and then executes instructional content. Evidence: The script skill_maker.py populates templates for SKILL.md files which are subsequently loaded by a sub-agent.
Audit Metadata