ads-generate
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONNO_CODE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted input from creative briefs and source assets, which is a potential surface for indirect prompt injection.
- Ingestion points: Creative brief text and source images loaded in Step 1 of SKILL.md.
- Boundary markers: Step 2 of SKILL.md explicitly instructs the agent to treat input as untrusted and ignore embedded instructions.
- Capability inventory: Step 6 of SKILL.md indicates the skill has the capability to write generated assets and manifests to the filesystem.
- Sanitization: Step 3 specifies building prompts from specific brand tokens and platform constraints rather than direct interpolation of the untrusted text.
- [NO_CODE]: No executable script files or source code were found in this skill, minimizing the direct attack surface.
Audit Metadata