leniu-redis-cache

Pass

Audited by Gen Agent Trust Hub on Mar 7, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: Indirect Prompt Injection Surface.
  • Ingestion points: Method parameters such as userId, taskId, and api paths documented in SKILL.md are used as components for building dynamic Redis keys.
  • Boundary markers: The documentation lacks explicit input validation or the use of specific delimiters to prevent key manipulation or command injection within the Redis context.
  • Capability inventory: The skill grants the agent the ability to perform Redis operations including data retrieval, insertion, pattern-based deletion, and acquiring distributed locks.
  • Sanitization: Code examples demonstrate direct string concatenation (e.g., 'user:' + userId) and string formatting without showing input filtering or sanitization of Redis-specific control characters or wildcards.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 7, 2026, 09:04 AM