klingai-pricing-basics

Pass

Audited by Gen Agent Trust Hub on Sep 9, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill's primary function involves analyzing 'workload descriptions' and 'synthetic prompts' provided by users to calculate costs, which represents a surface for instructions to be injected into the agent context.\n
  • Ingestion points: SKILL.md instructions direct the agent to process workload data and canary prompts.\n
  • Boundary markers: The skill does not utilize delimiters or explicit instructions to ignore potentially malicious embedded content within the workload data.\n
  • Capability inventory: The skill has access to sensitive tools including Bash(npm:*), Read, Write, Edit, and Grep.\n
  • Sanitization: There is no implemented logic to sanitize or validate the external data before it is processed by the agent.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 9, 2026, 03:48 AM
Security Audit — agent-trust-hub — klingai-pricing-basics