The Agent Skills Directory

[SAFE]: The skill is primarily instructional documentation and does not contain executable code, remote downloads, or obfuscated content. It explicitly mandates security best practices for developers, such as:
Using environment variables for sensitive data rather than hardcoding secrets.
Implementing parameterized queries for database operations to prevent injection attacks.
Enforcing strict input validation for all external data sources.
Applying thread synchronization (locks) to prevent race conditions in critical trading operations.
[INDIRECT_PROMPT_INJECTION]: The skill describes an architecture that processes untrusted external data (market data from Polymarket/Binance and notifications). This constitutes an indirect prompt injection attack surface.
Ingestion points: Data fetching in src.data.market_data and notification processing logic.
Boundary markers: The skill does not explicitly define delimiters for separating external content from agent instructions.
Capability inventory: The bot has capabilities to write to a database (db_connection) and execute trades (execute_trade).
Sanitization: The skill mitigates these risks by requiring "proper input validation for all external data" and "robust parsing" for notifications in its Security Standards section.

python-bot-standards