skills/jmsktm/claude-settings/llc-ops/Gen Agent Trust Hub

llc-ops

Pass

Audited by Gen Agent Trust Hub on Feb 16, 2026

Risk Level: LOW
Full Analysis
  • PROMPT_INJECTION (SAFE): No malicious override or bypass instructions were detected. The use of specialized agent roles and invocation patterns are legitimate structural instructions for multi-agent orchestration.
  • DATA_EXFILTRATION (SAFE): The skill does not contain any code for network communication, file reading from sensitive system paths, or credential harvesting. Placeholders for EIN and SSN are for user reference only.
  • REMOTE_CODE_EXECUTION (SAFE): There are no scripts, package manager commands, or remote download patterns. The skill is entirely composed of static Markdown documentation.
  • INDIRECT_PROMPT_INJECTION (LOW): While the agent is designed to process external documents like audit notices or expense receipts, it lacks tool-calling capabilities or side-effect-heavy actions. The risk is limited to reasoning logic.
  • OBFUSCATION (SAFE): No encoded strings, homoglyphs, or hidden Unicode characters were identified across the 6 files.
Audit Metadata
Risk Level
LOW
Analyzed
Feb 16, 2026, 05:00 AM