dataforseo
Pass
Audited by Gen Agent Trust Hub on Aug 24, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill processes untrusted external data retrieved from search engines (Google SERPs, business listings) and AI search results (ChatGPT, Claude, Gemini, Perplexity responses). This represents a potential surface for indirect prompt injection attacks where malicious instructions could be embedded in search snippets or AI answers to influence the agent's behavior.
- Ingestion Points: Data is ingested through the
dataforseoLiveanddataforseoLiveRawhelpers inlib/dataforseo.ts, which feed results to 33 individual tool scripts. - Boundary Markers: The skill does not explicitly instruct the agent to ignore or delimit instructions found within the retrieved search data in its tool definitions or
SKILL.mdguidance. - Capability Inventory: The tools themselves are read-only and restricted to specific API endpoints. The risk is limited to the agent's downstream interpretation of the data rather than direct capability abuse by the skill itself.
- Sanitization: While output is validated against Zod schemas, the textual content of search results (titles, descriptions, HTML) is not sanitized for LLM instructions.
Audit Metadata