The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from several external websites to perform its analysis. This creates a vulnerability surface where malicious content on these pages could influence the agent's behavior.
Ingestion points: Matchup data from Yahoo Fantasy Sports (baseball.fantasysports.yahoo.com), schedules from MLB.com, and pitcher data from RotoWire and FanGraphs.
Boundary markers: The skill does not explicitly define delimiters for untrusted data in its processing instructions, though it does emit data in a structured YAML format.
Capability inventory: The skill writes to the local file system (signals/ directory) and delegates logic to other agent skills.
Sanitization: The workflow includes a validation step using an internal tool mlb-signal-emitter before file persistence.
[EXTERNAL_DOWNLOADS]: The skill routinely fetches data from well-known sports and technology services. These downloads are performed for legitimate skill functionality.
Evidence: Fetches matchup scores from Yahoo, schedules from MLB.com, and player statistics from FanGraphs and RotoWire as described in SKILL.md and resources/methodology.md.

mlb-category-state-analyzer