perplexity-core-workflow-b
Pass
Audited by Gen Agent Trust Hub on Aug 26, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill provides templates and logic for research workflows using the Perplexity API. It correctly uses environment variables for the API key and interacts only with official API endpoints.
- [INDIRECT_PROMPT_INJECTION]: The skill processes research results from the Perplexity API and compiles them into a structured Markdown document. This creates a surface for indirect prompt injection where malicious instructions embedded in external web content could be passed to the agent. 1. Ingestion points: The
researchTopicandconductResearchfunctions ingest external data from API responses. 2. Boundary markers: The compiled Markdown output lacks explicit delimiters or warnings to the agent that the content is derived from untrusted external sources. 3. Capability inventory: The skill utilizesWriteandEdittools to generate research reports based on external data. 4. Sanitization: No sanitization or filtering of the API-provided answers is performed before interpolation into the final report. As this behavior is associated with the primary purpose of the skill and no active exploitation was found, it does not escalate the verdict.
Audit Metadata