covered-call-cash-secured-put-income
Pass
Audited by Gen Agent Trust Hub on Sep 3, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from an external source, creating a potential surface for indirect prompt injection.\n
- Ingestion points: The script
scripts/covered_call_cash_secured_put_income_diagnostics.pyreads data from a user-provided file namedinput.csvviapd.read_csv.\n - Boundary markers: No explicit delimiters or instructions are provided to the agent to ignore potentially malicious content within the CSV file or its headers.\n
- Capability inventory: The script performs file read/write operations and provides numeric summaries to the agent. No network or subprocess capabilities are present in the script.\n
- Sanitization: The script coerces values in data columns to numeric or datetime types, which effectively strips malicious strings from the data rows. However, column names (headers) from the CSV are extracted and returned in the JSON summary as a list of strings (
summary[\"columns\"]) without sanitization, allowing an attacker to inject text into the agent's context through malformed headers.
Audit Metadata