LLMSkillHub
mirroredCopied from OctagonAI/octagon-mcp-server under MIT. OctagonAI has not claimed this page, and is not affiliated with LLM SkillHub.

octagonai / octagon-api-smoke-test

Validate Octagon plugin configuration and run a lightweight smoke test across the main Octagon MCP workflows. Use when checking whether the plugin is configured correctly, debugging auth, or verifying tool availability.

0
SKILL.md
1.5 KB

Octagon API Smoke Test

Use this skill to confirm that the Claude plugin and Octagon MCP are configured correctly.

Smoke test sequence

  1. Run a simple octagon-agent request
  2. Run a simple octagon-deep-research-agent request
  3. If a Kalshi URL is available, run octagon-prediction-markets-agent
  4. If an event ticker is available, run prediction_markets_history

MCP calls

General agent:

{
  "server": "octagon-mcp",
  "toolName": "octagon-agent",
  "arguments": {
    "prompt": "Give me a one-sentence summary of Apple's latest quarter."
  }
}

Deep research:

{
  "server": "octagon-mcp",
  "toolName": "octagon-deep-research-agent",
  "arguments": {
    "prompt": "Research the current AI infrastructure spending cycle in one short paragraph."
  }
}

Failure triage

  • Missing auth or invalid key: check plugin api_key
  • Unexpected environment or staging issue: check api_base_url
  • Prediction market failures mentioning Kalshi URL: supply a valid Kalshi market URL
  • Credit or entitlement failures: report the exact error and stop

Success criteria

  • At least one Octagon tool returns usable text
  • Tool failures, if any, are classified clearly as config, entitlement, input, or service issues