The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill instructions define an agent that processes untrusted data from multiple external sources, creating an attack surface where embedded instructions could influence the agent's behavior or analytical output. \n- Ingestion points: The skill leverages SEC filings, earnings transcripts, news, social media sentiment, web traffic data, app store data, and patent filings as mentioned in SKILL.md. \n- Boundary markers: The instructions do not specify any delimiters or explicit 'ignore' directives to prompt the agent to disregard instructions found within external content. \n- Capability inventory: The skill documentation indicates the usage of Python (including pandas, numpy, statsmodels, and yfinance) and R for quantitative and fundamental analysis of this data. \n- Sanitization: No sanitization, filtering, or validation steps for untrusted external content are outlined in the skill methodology.

Investment Researcher