kpi-review

Pass

Audited by Gen Agent Trust Hub on Sep 4, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data from user-uploaded CSV files or pasted text without explicit sanitization or boundary markers. This creates a potential surface for indirect prompt injection if the input data contains malicious instructions aimed at influencing the agent's behavior.
  • Ingestion points: The skill is instructed to parse CSV reports from the uploads/ directory or numbers provided in the chat interface as primary data sources (SKILL.md).
  • Boundary markers: The instructions do not define specific delimiters or guardrails to treat the external data strictly as data rather than instructions.
  • Capability inventory: The skill has the capability to write to the local filesystem, specifically creating files in output/kpi-review/ and appending data to knowledge/kpis.md (SKILL.md).
  • Sanitization: There are no documented steps for validating, filtering, or sanitizing the input data before it is processed by the agent.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 4, 2026, 06:28 AM
Security Audit — agent-trust-hub — kpi-review