sentiment-analysis
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and synthesize untrusted data from external feedback sources, including CSV files, PDFs, and survey responses. This pattern creates a potential surface for indirect prompt injection, where an attacker could embed instructions within the feedback data to manipulate the agent's summary or recommendations.
- Ingestion points: Reads CSV, PDF, survey responses, and review data (SKILL.md).
- Boundary markers: No explicit delimiters or instructions to ignore embedded commands were identified.
- Capability inventory: The skill instructions focus on analytical tasks and synthesis; no dangerous system commands or network operations are explicitly requested.
- Sanitization: There are no instructions for sanitizing or filtering input data before processing.
Audit Metadata