stock-trade-journal

Warn

Audited by Gen Agent Trust Hub on Mar 10, 2026

Risk Level: MEDIUMDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [DATA_EXFILTRATION]: The skill is vulnerable to path traversal in scripts/record_trade.py. The --ts-code parameter is used to build the Markdown file path using os.path.join(base, "records", f"{args.ts_code}.md"). Without validation, this allows writing files to arbitrary locations by providing absolute paths or using directory traversal sequences.\n- [PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection (Category 8).\n
  • Ingestion points: Untrusted data enters the agent context through command-line arguments in scripts/record_trade.py (specifically the --reason, --note, and --take-profit flags) and scripts/query_trades.py (via the --ts-code flag).\n
  • Boundary markers: The skill does not implement delimiters or warnings to ignore embedded instructions within the processed data fields.\n
  • Capability inventory: The skill has the capability to write to Markdown files and perform SQLite database operations.\n
  • Sanitization: While the skill correctly uses parameterized queries for SQLite database operations, it lacks sanitization or escaping for user-provided strings before they are appended to Markdown files in scripts/record_trade.py.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Mar 10, 2026, 03:26 AM