The Agent Skills Directory

[SAFE]: No malicious patterns, obfuscation, or unauthorized data access were found in the provided files. The skill emphasizes manual verification and rigorous validation of trading logic.- [COMMAND_EXECUTION]: The skill references the execution of local Python scripts for data collection (scripts/research_dataset.py) and backtesting. These are standard operations for the described quantitative research workflow within the agent's environment.- [INDIRECT_PROMPT_INJECTION]: The skill defines a protocol for processing external trading ideas by converting them into structured, human-readable rules in RULES.md. This practice acts as a security boundary, ensuring that external instructions are reviewed and frozen before any code is generated or executed.

desk-strategy-lab