The Agent Skills Directory

[PROMPT_INJECTION]: The skill has a surface for indirect prompt injection as it requires the agent to ingest and process untrusted data from the web.
Ingestion points: The agent is instructed to perform WebSearch queries for market themes and evaluate the findings to adjust confidence levels (SKILL.md, Step 4).
Boundary markers: Absent. The instructions do not define delimiters or specific markers to separate external search content from the agent's internal reasoning.
Capability inventory: The skill executes local Python scripts (theme_detector.py) and writes files to the reports/ directory.
Sanitization: Absent. External narrative data is evaluated directly by the model without filtering.
[EXTERNAL_DOWNLOADS]: The skill intentionally fetches market data from established financial services.
Fetches market and stock metrics from the Financial Modeling Prep (FMP) and FINVIZ APIs to populate theme scores.
Downloads historical sector trend data in CSV format from a public GitHub repository (tradermonty/uptrend-dashboard).
These network operations are strictly for data acquisition and are standard for the skill's purpose.

theme-detector