market-research

Pass

Audited by Gen Agent Trust Hub on Mar 12, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection due to its core functionality of ingesting and processing untrusted data from web search results.
  • Ingestion points: External data is retrieved via web search commands in files step-02-customer-behavior.md through step-06-research-completion.md.
  • Boundary markers: The workflow lacks explicit boundary markers or system-level instructions to treat search results strictly as data rather than instructions.
  • Capability inventory: The skill has the capability to write to the local filesystem (the {planning_artifacts} path) and is instructed to use subagents or subprocesses for parallel analysis.
  • Sanitization: There is no evidence of sanitization, filtering, or validation of the retrieved web content before it is processed by the agent into the final report.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 12, 2026, 09:05 PM