ads-create
Pass
Audited by Gen Agent Trust Hub on Sep 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes data from local files (
brand-profile.json,ADS-AUDIT-REPORT.md, and*-audit-results.md) and interpolates their content into instructions for sub-agents. This creates a surface for indirect prompt injection if the files contain malicious instructions. - Ingestion points:
brand-profile.jsonand audit reports are loaded in SKILL.md. - Boundary markers: Absent; the skill does not use delimiters or explicit warnings to isolate untrusted file content.
- Capability inventory: Spawning sub-agents (
creative-strategist,copy-writer) and writing to the file system (campaign-brief.md). - Sanitization: Absent; content is used directly to guide agent behavior.
- [DYNAMIC_EXECUTION]: The skill orchestrates a multi-agent workflow by spawning sub-agents and providing them with instructions derived from local file content. This reliance on external file integrity for agent behavior is a form of dynamic instruction generation.
- Execution method: Sequential spawning of
creative-strategistandcopy-writerusing the platform's task tool.
Audit Metadata