Prompts Reference

MCP prompts are guided workflow templates that structure multi-step interactions. The LLM client receives the prompt text and follows the workflow, calling the appropriate tools at each step. All four prompts are registered on the mcnoaa-tides server.

Tip

Prompts are not tools — they do not execute logic or return data directly. They provide structured instructions that guide the LLM through a sequence of tool calls to accomplish a task.

---

plan_fishing_trip

Guided fishing trip planning using tide and weather data. Walks through station discovery, tide analysis, and conditions assessment to recommend optimal fishing windows.

Parameter	Type	Required	Default	Description
location	str	yes	—	Place name or area description (e.g. "Narragansett Bay")
target_species	str	no	""	Target species for species-specific recommendations
date	str	no	""	Planned date (defaults to today if empty)

Workflow

1. Station discovery — Uses find_nearest_stations to locate NOAA tide stations near the specified location and selects the closest station with tide prediction support.

2. Tide analysis — Uses get_tide_predictions to find the next high and low tides. Incoming (rising) tide is generally better for most species as it pushes bait and forage into estuaries and along structure.

3. Conditions check — Uses marine_conditions_snapshot to assess current conditions:

   • Wind: sustained > 15 kn makes small-boat fishing uncomfortable
   • Water temperature: affects species activity and feeding patterns
   • Pressure trend: falling pressure often improves bite rates

4. Window recommendation — Synthesizes a fishing window that aligns favorable tide phase with manageable weather, including expected water level, weather summary, and safety concerns.

Example usage

{
  "prompt": "plan_fishing_trip",
  "arguments": {
    "location": "Puget Sound near Seattle",
    "target_species": "chinook salmon",
    "date": "2026-02-25"
  }
}

The LLM receives the full workflow template and executes each step, calling find_nearest_stations, get_tide_predictions, and marine_conditions_snapshot in sequence.

---

marine_safety_check

Go/no-go safety assessment for a marine station. Evaluates current conditions against common safety thresholds for recreational boating and fishing.

Parameter	Type	Required	Default	Description
station_id	str	yes	—	7-digit NOAA station ID (e.g. "8454000")

Assessment criteria

The prompt instructs the LLM to evaluate conditions against these thresholds. If any single factor is NO-GO, the overall assessment should be NO-GO.

Wind

Assessment	Sustained	Gusts
GO	< 15 kn	< 20 kn
CAUTION	15-20 kn	20-25 kn
NO-GO	> 20 kn	> 25 kn

Visibility

Assessment	Range
GO	> 2 nm
CAUTION	1-2 nm
NO-GO	< 1 nm

Water temperature (hypothermia risk)

Risk Level	Temperature
LOW RISK	> 70 F
MODERATE	60-70 F (wear PFD, limit exposure)
HIGH RISK	< 60 F (survival suit or dry suit recommended)

Pressure trend

Rapidly falling pressure (> 3 mb over 3 hours) suggests an approaching storm system.

Workflow

1. Data fetch — Uses marine_conditions_snapshot to retrieve all available conditions for the station.

2. Threshold evaluation — Evaluates each condition (wind, visibility, water temperature, pressure trend) against the criteria above.

3. Assessment — Provides a clear GO / CAUTION / NO-GO recommendation with reasoning for each factor.

Example usage

{
  "prompt": "marine_safety_check",
  "arguments": {
    "station_id": "9447130"
  }
}

---

smartpot_deployment

Full SmartPot deployment planning workflow. Orchestrates station discovery, tidal phase analysis, deployment briefing, and anomaly detection into a single deployment recommendation.

Parameter	Type	Required	Default	Description
latitude	str	yes	—	Deployment latitude (passed as string)
longitude	str	yes	—	Deployment longitude (passed as string)
soak_hours	str	no	"48"	Planned soak duration in hours (passed as string)

Note

Parameters are strings because MCP prompt arguments are always strings. The prompt template passes these values into tool calls where they are parsed to the correct types.

Workflow

1. Station discovery — Uses find_nearest_stations to identify the reference tide station. Notes the distance, since data accuracy decreases beyond ~10 nm from the station.

2. Tidal phase — Uses tidal_phase to determine the current phase. Deploying during slack or early flood maximizes initial soak productivity.

3. Deployment briefing — Uses deployment_briefing to get the GO/CAUTION/NO-GO assessment, including tide schedule, wind, temperature, and pressure advisories for the soak window.

4. Anomaly check — Uses water_level_anomaly on the nearest station to detect storm surge or unusual water level deviations that could affect pot stability or recovery.

5. Synthesis — Produces a deployment recommendation with:

   • Overall GO/CAUTION/NO-GO with reasoning
   • Optimal deployment window within the next 6 hours
   • Expected tidal cycles during soak
   • Safety or recovery concerns
   • Recommended recovery time based on tide schedule

Example usage

{
  "prompt": "smartpot_deployment",
  "arguments": {
    "latitude": "47.6",
    "longitude": "-122.34",
    "soak_hours": "72"
  }
}

---

crab_pot_analysis

Guided catch-data analysis with tidal correlation. Helps analyze historical catch events against tidal phases to identify productive patterns.

Parameter	Type	Required	Default	Description
events_description	str	no	""	Description of catch data context or data source

Workflow

1. Data gathering — Collects catch event data. Each event needs at minimum: timestamp (UTC), latitude, longitude, and catch_count or weight. Additional fields like species, pot_id, and bait_type are preserved.

2. Tidal enrichment — Uses catch_tidal_context to classify each event as ebb, flood, slack_high, or slack_low with progress through the phase.

3. Phase aggregation — Groups enriched events by tidal phase and computes average catch per phase, best-performing phase, and catch distribution.

4. Pattern analysis — Examines whether certain phases consistently outperform, whether there is a sweet spot in phase progress (e.g., early flood vs late flood), and any correlation with tide range (spring vs neap).

5. Summary — Produces a report with top-performing tidal conditions, recommended deployment timing relative to tide, confidence level based on sample size, and suggestions for future data collection.

Example usage

{
  "prompt": "crab_pot_analysis",
  "arguments": {
    "events_description": "Three weeks of Dungeness crab catch data from Puget Sound, 6 pots"
  }
}

The LLM will prompt for the actual event data (or read it from a provided source), then run the analysis workflow.