ads-generate
Pass
Audited by Gen Agent Trust Hub on Sep 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted input from campaign-brief.md and brand-profile.json to generate prompts for a spawned sub-agent (visual-designer). This creates a vulnerability where instructions embedded in the brief could influence the sub-agent's behavior.\n
- Ingestion points: campaign-brief.md and brand-profile.json (SKILL.md).\n
- Boundary markers: None; the skill lacks instructions to delimit the brief content or warn the sub-agent about potential embedded instructions.\n
- Capability inventory: Spawns sub-agents with context: fork, writes files to the local directory structure (./ad-assets/), and reads configuration from ~/.banana/ and ~/.claude/ directories.\n
- Sanitization: None; the brief content is parsed directly without validation or filtering.\n- [COMMAND_EXECUTION]: The skill references the use of a fallback script scripts/generate_image.py and the execution of the /banana tool for setup and generation tasks.
Audit Metadata