Loading ToolBench

Reading the current corpus…

poke-power-outage-monitoring MCP server · F · ToolBench

poke-power-outage-monitoring

github.com/johnyrokita/poke-power-outage-monitoring

MCP server for monitoring power outages in a specified location by querying the PGE (Polish power company) API and providing real-time outage status information

<=2025-11-25STREAMABLE_HTTP1 tools

34/100

Definition Quality25

Protocol Readiness69

Supportability26

Spec posture

Inferred effective spec <=2025-11-25·current baseline 2026-07-28

Static source inference · medium confidence · detected: stateful session

Deprecated protocol patterns detected

Summary

Single tool 'get_power_outages' has critical definition gaps. The tool description mentions returning 'outage information including start/stop times, duration, and affected addresses' but actual output schema is undocumented, LLM cannot plan downstream calls or extract structured data. Error handling returns plain text with isError flag but provides no recovery guidance. No tool annotations (readOnlyHint, destructiveHint, idempotentHint). Naming follows verb_noun pattern (get_power_outages), acceptable but generic.

Tools (1)

get_power_outagesread onlysource verified25/100

Get current power outage status for location: ${LOCATION}. Returns outage information including start/stop times, duration, and affected addresses.

Issues (5)

CRITICALconstrained-input ↗

Input schema is empty ({}). Tool accepts no documented parameters, yet description references 'location: ${LOCATION}', location is hardcoded server-side, not a parameter. This breaks the tool's flexibility and prevents agents from querying different locations.

get_power_outages

CRITICALtool ↗

Output schema is not documented. Tool returns JSON-stringified status object with fields like 'hasOutage', 'outages' (array with id, startAt, stopAt, affectedAddresses), and 'lastChecked', but LLM has no schema to understand structure. Prevents proper data extraction and downstream tool composition.

get_power_outages

CRITICALtool-description ↗

No parameter descriptions. Tool has zero input parameters documented, making it impossible for LLMs to understand what inputs (if any) are valid or required.

get_power_outages

HIGHrecovery-guide ↗

Error handling lacks recovery guidance. When getOutageStatus() fails, tool returns plain error message with isError=true, but provides no actionable next steps (e.g., 'Check POWER_OUTAGE_LOCATION env var' or 'Retry in 30 seconds').

Recommendations

Add 'location' as a required string parameter with description: 'The geographic location to query for power outages (e.g., "San Francisco, CA"). If not provided, defaults to POWER_OUTAGE_LOCATION env var.' This allows agents to query multiple locations.
Document output schema explicitly: 'Returns object with fields: hasOutage (boolean), outages (array of {id: number, startAt: ISO8601, stopAt: ISO8601, affectedAddresses: string[]}), lastChecked (ISO8601).'
Enhance error handling: return structured error with actionable guidance. E.g., 'Error: Failed to fetch outage status. Cause: [specific error]. Next steps: Verify POWER_OUTAGE_LOCATION is set correctly, or retry in 30 seconds.'
Add readOnlyHint annotation to tool definition to signal to agents that this tool is safe to call without side effects.
Expand tool description to clarify when to use it: 'Call this tool to check current power outage status for a location. Returns active outages with start/stop times and affected addresses. Use before notifying users or triggering alerts.'

Spec posture evidence

Inferred effective spec: <=2025-11-25.

Relies on Stateful initialize / Mcp-Session-Id (removed; protocol is stateless) - make each request self-contained

Score history

Overall score trend

First recorded score · v2 rubric

34/100

Scored	Grade	Overall	Spec posture	Rubric
2026-09-23	F	34	<=2025-11-25	v2