ask-ai-bookkeeping

Pass

Audited by Gen Agent Trust Hub on Aug 18, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructions provide a shell command template: python3 "$AI_BOOKKEEPING_HOME/scripts/bookkeeping.py" --config "/absolute/path/to/config.json" ask "用户问题". Passing raw user input ("用户问题") directly into a shell command is a standard command injection vector if the execution environment does not perform proper escaping or use safe argument passing mechanisms.
  • [INDIRECT_PROMPT_INJECTION]: The skill defines an attack surface for indirect prompt injection by taking untrusted natural language input from the user and using it as a parameter for a powerful capability (executing a Python script that reads a ledger).
  • Ingestion points: User queries provided as the "用户问题" argument in the query flow in SKILL.md.
  • Boundary markers: None specified to prevent the agent from misinterpreting embedded instructions in the user's question.
  • Capability inventory: Execution of a local Python script via python3 (bookkeeping.py) which has access to the filesystem (ledger data and configuration file).
  • Sanitization: No explicit instructions are provided to sanitize or escape the user-provided string before passing it to the shell command.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 18, 2026, 04:18 PM
Security Audit — agent-trust-hub — ask-ai-bookkeeping