specialist-marketing-book-co-author
Pass
Audited by Gen Agent Trust Hub on Aug 15, 2026
Risk Level: SAFEPROMPT_INJECTIONNO_CODE
Full Analysis
- [PROMPT_INJECTION]: The skill takes user input directly through the
{the user's question}placeholder and interpolates it into the system prompt of a spawned sub-agent. There are no boundary markers (e.g., XML tags or delimiters) or specific instructions to ignore embedded commands, which could allow a user to override the sub-agent's persona or instructions. - [INDIRECT_PROMPT_INJECTION]: The skill possesses a vulnerability surface for indirect prompt injection.
- Ingestion points: User-provided input is accepted via the placeholder in
SKILL.md. - Boundary markers: None are present to delimit the user's question from the agent's instructions.
- Capability inventory: The skill uses the
Agenttool to spawn sub-agents. - Sanitization: There is no evidence of input validation or escaping for the user-provided data.
- [NO_CODE]: The skill consists entirely of markdown and instructions for the agent platform; it does not include any executable scripts (Python, JavaScript, etc.) or configuration files that could hide malicious logic.
Audit Metadata