Content Repurposing
Pass
Audited by Gen Agent Trust Hub on Aug 23, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data in Step 1, where it requires a link, transcript, or draft of a source asset. The instructions lack boundary markers or sanitization logic to distinguish between the source content and potential instructions embedded within that content. While the skill's capabilities are limited to generating marketing copy and distribution plans, an attacker could potentially influence the agent's output by embedding malicious instructions in the provided source material.
- Ingestion points: File
SKILL.md(Step 1: gather inputs - source asset).
- Boundary markers: Absent; no instructions provided to treat source text as data rather than instructions.
- Capability inventory: Text generation and content mapping; no high-privilege tool usage detected.
- Sanitization: Absent; no validation or filtering of the source asset is defined.
Audit Metadata