The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill retrieves market data from several well-known and official financial services. These include the U.S. Department of the Treasury (interest rates), the Federal Reserve Bank of St. Louis (FRED economic data), Cboe (VIX volatility index), the European Central Bank (FX reference rates), and the Bank of Korea (ECOS market indicators). These sources are recognized as highly reputable and the communication is handled via a local Python script.
[COMMAND_EXECUTION]: The skill workflow involves executing a bundled Python helper script (scripts/market_data.py) to perform data collection and normalization. The agent is also instructed to use a CLI tool for setup and instruction retrieval. The local script uses standard library modules and includes safety mechanisms such as response size limits and timeout handling.
[INDIRECT_PROMPT_INJECTION]: The skill processes external financial data to generate market summaries, which introduces an indirect prompt injection attack surface. However, the risk is mitigated by the skill's architecture and strict validation rules.
Ingestion points: Market data payloads (CSV and JSON) retrieved by scripts/market_data.py from external financial APIs.
Boundary markers: The skill provides a rigorous validation framework in references/format-rules.md, requiring the agent to verify session dates, check for data conflicts, and use multiple independent sources to cross-verify level Level Level Level Level Level findings.
Capability inventory: The script is limited to HTTP GET requests using standard libraries; the agent's actions are restricted to reading structured script output and generating text briefings. No arbitrary code execution or filesystem write capabilities are exposed to the ingested data.
Sanitization: The helper script strictly parses external payloads into typed numeric values (floats) before they are passed to the agent, preventing raw, potentially malicious strings from entering the context.

multi-asset-morning-briefing