Skanetrafiken Trip Planner

**Overall judgment:** High-quality, mature trip-planning skill focused on Skånetrafiken, with strong LLM-oriented documentation. Well-suited as the primary public transport planner for Skåne and cross-border journeys to Copenhagen. **Strengths** - Clear, minimal API surface (`search-location.sh` and `journey.sh`) that maps well to typical agent workflows. - Extensive instructions for interpreting user time intent (now / depart later / arrive by / around / relative times) and converting to concrete query datetimes. - Robust handling of different location types (STOP_AREA, ADDRESS, POI, LOCATION) with explicit conversion rules to satisfy the Journey API’s stricter requirements. - Highly user-friendly response-format guidance (emojis, walking distances, walking-time estimation, transfers, delays) that makes it easy for an LLM to present readable itineraries. - Error-handling patterns explicitly described for ambiguous locations, no results, and cross-border naming issues. **Key risks / limitations** - **Geographic scope:** Primarily limited to Skåne and common Copenhagen destinations; not a general-purpose global transport planner. - **External dependency:** Reliability and latency are tied to Skånetrafiken’s backend and network stability; no explicit retry/backoff strategy is described. - **Search sensitivity:** POI and station search is formatting-sensitive (e.g., “Emporia” vs “Emporia, Malmö”; correct “C” station names required). An agent that ignores these rules may see degraded results. - **Time semantics:** The skill expects the LLM to correctly parse and normalize relative and vague time expressions into explicit datetimes; misinterpretation here can lead to suboptimal journey suggestions. **Recommended scenarios** - Assistants aimed at residents or visitors in Skåne/Copenhagen who frequently use public transport. - Travel-planning chatbots that need to reason about different departure options, transfers, and delays in real time. - Multimodal city guides where the LLM must convert free-form location and time queries into precise, actionable transit instructions.