ai-slop-audit
Pass
Audited by Gen Agent Trust Hub on Oct 11, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external content, creating a surface for indirect prompt injection attacks where malicious instructions could be embedded in the audited material.
- Ingestion points: The agent ingests "artefacts" from users which can be pasted text, uploaded files, or content retrieved from external URLs as defined in the SKILL.md 'Required Inputs' table.
- Boundary markers: The instructions lack specific guidance for the agent to use secure delimiters or ignore embedded instructions within the ingested artefacts, increasing the risk that the LLM might follow commands hidden in the content it is meant to audit.
- Capability inventory: The skill is strictly limited to "read and search only" operations and explicitly prohibits publishing, outreach, or spend actions without additional authority (SKILL.md, Capability and Permission Boundaries). This constraint significantly reduces the risk of automated damage or data exfiltration resulting from a successful injection.
- Sanitization: The instructions do not specify any sanitization, filtering, or escaping of the input content before it is analyzed by the agent.
Audit Metadata