report server

Submit an agent usage report for an MCP server. Reports are aggregated and influence registry trust scores. Call this after using an MCP server so the registry

liveverifiedunknown~5s typical

mcpprobedarclanmcpregistry

$0.0020

per call

plus $0.001 platform fee per run

Open MCP server; Cracked bills routing only

Provider Upstream docs

$Set up https://cracked-ai-pearl.vercel.app/SKILL.md, then use Cracked to run mcp-ai-arclan-registry /report_server for me.

Input

Field	Type	Description
serverUrl*	string	The MCP server endpoint URL that was used
outcome*	string (success \| failure \| partial)	Result of using the server
agentId	string	Identifier for the reporting agent (optional)
toolsUsed	array	List of tool names that were called
turnsRequired	number	Number of turns to complete the task
latencyMs	number	Total wall-clock latency in milliseconds
taskCategory	string	Category of task attempted (e.g. "search", "write", "read")
taskDescription	string	Brief description of what the task was
errors	array	Array of error strings encountered during tool use
criterion	string (output_reliability \| side_effect_transparency \| failure_handling \| rate_limits \| agent_feedback)	Specific reliability criterion being scored (enables structured component scoring)
componentScore	number	Score 0–100 for the specified criterion
severity	string (info \| warning \| error \| critical)	Severity of the worst issue encountered
environment	string (prod \| staging \| dev \| unknown)	Environment the server was used in
evidenceRef	string	Trace ID, log URL, or other reference supporting this report

Call it

curl

curl https://cracked-ai-pearl.vercel.app/v1/run \
  -H "Authorization: Bearer ck_live_..." -H "content-type: application/json" \
  -d '{"provider":"mcp-ai-arclan-registry","endpoint":"/report_server","input":{"serverUrl":"...","outcome":"success"}}'

cli

npx cracked-ai run -p mcp-ai-arclan-registry -e /report_server -i '{"serverUrl":"...","outcome":"success"}'

mcp

run_tool({ provider: "mcp-ai-arclan-registry", endpoint: "/report_server", input: {"serverUrl":"...","outcome":"success"} })

Try it

Runs against your signed-in workspace balance. Sign in if you have not.