record-ai-bookkeeping

Warn

Audited by Gen Agent Trust Hub on Aug 18, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill provides templates for executing shell commands using user-supplied strings, such as python3 "$AI_BOOKKEEPING_HOME/scripts/bookkeeping.py" ... record "用户原话". This implementation is vulnerable to command injection if the user input contains shell metacharacters (e.g., semicolons, backticks, or dollar-sign syntax) that are not properly escaped or sanitized by the agent platform before execution.
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted data from the user to drive transaction logging and script execution.
  • Ingestion points: Natural language input describing transaction details (amount, account, category, etc.) provided by the user in the prompt.
  • Boundary markers: The instructions use double quotes to wrap the user input variable in shell command examples, which does not provide security against nested quotes or shell escapes within the user's string.
  • Capability inventory: The skill possesses the capability to execute a Python script (bookkeeping.py) which can read/write configuration files, modify databases (record/update/undo transactions), and handle environment variables.
  • Sanitization: The instructions lack specific guidance or requirements for sanitizing, validating, or escaping the user's raw input before it is passed to the shell environment.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Aug 18, 2026, 04:18 PM
Security Audit — agent-trust-hub — record-ai-bookkeeping