data-algo-fin

Warn

Audited by Gen Agent Trust Hub on Jun 17, 2026

Risk Level: MEDIUMDATA_EXFILTRATIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [DATA_EXFILTRATION]: The skill references absolute local file paths (e.g., /Users/0xvox/.../tv_webhook.py) and project-specific memory paths. This exposes information about the host system's user directory structure.
  • [COMMAND_EXECUTION]: The skill is designed to modify the local environment by writing Python modules into the propfirm_engine/ directory and strategy profiles to the .strategy-profile/ folder. It also involves running local backtest scripts (propfirm_engine/backtest.py), which executes the generated or modified code.
  • [PROMPT_INJECTION]: The skill has an indirect prompt injection surface:
  • Ingestion points: The agent reads external "journal data" and "strategy code" provided by the operator to diagnose weaknesses.
  • Boundary markers: Absent. There are no instructions to use delimiters or to ignore instructions embedded in the processed data.
  • Capability inventory: The agent can write executable files to the local filesystem and trigger script execution for backtesting.
  • Sanitization: Absent. The skill does not perform validation or sanitization of the input data before using it to generate implementation code.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Jun 17, 2026, 12:49 PM
Security Audit — agent-trust-hub — data-algo-fin