thesis-drift
Warn
Audited by Gen Agent Trust Hub on Jul 4, 2026
Risk Level: MEDIUMCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill automates the execution of shell commands using the tools/financial_rigor.py script. It populates command arguments (such as --price, --eps, and --bvps) with data extracted directly from user-provided or discovered investment reports. This interpolation of untrusted data into shell commands without sanitization creates a risk of command injection, where a specially crafted report could execute unauthorized commands on the underlying system.
- [PROMPT_INJECTION]: The skill processes external, untrusted content from investment reports without adequate safety boundaries. This exposes the agent to indirect prompt injection, where malicious instructions hidden in a report could override the skill's intended logic or hijack the agent's session.
- Ingestion points: Investment reports and historical snapshots located in the reports/ directory.
- Boundary markers: Absent. There are no explicit delimiters or safety instructions provided to the agent to isolate report content from the skill's execution logic.
- Capability inventory: The skill utilizes shell command execution, file system access, and the execution of local scripts.
- Sanitization: Absent. Extracted data is used verbatim in subsequent analytical and computational steps.
Audit Metadata