ib-stop-loss
Pass
Audited by Gen Agent Trust Hub on Sep 19, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes a local Python script (
scripts/stop_loss.py) usinguv runto perform brokerage operations. This is the intended behavior for managing trading positions and generating reports. - [INDIRECT_PROMPT_INJECTION]: The skill ingests portfolio and market data (symbols, strikes, expiry, prices) from the Interactive Brokers API to generate markdown reports for the user.
- Ingestion points: Data is fetched from the Interactive Brokers API through the
get_stop_loss_datafunction inscripts/stop_loss.py. - Boundary markers: The report instructions in
SKILL.mduse standard markdown formatting (tables and headers) but lack explicit delimiters or "ignore embedded instructions" warnings to isolate potentially untrusted data. - Capability inventory: The skill possesses the capability to place, modify, and cancel trading orders (via the IBKR API) and perform network communication with the local IBKR Gateway/TWS when the
--executeflag is provided. - Sanitization: While the internal script outputs structured JSON, the subsequent markdown report generation does not explicitly demonstrate sanitization for content that could theoretically be present in account-level metadata or custom position names.
Audit Metadata