octagon-mcp
Version:
MCP server for Octagon API. Provides specialized AI agents for investment research of public and private markets.
54 lines (40 loc) • 1.52 kB
Markdown
name: octagon-api-smoke-test
description: 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.
# 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:
```json
{
"server": "octagon-mcp",
"toolName": "octagon-agent",
"arguments": {
"prompt": "Give me a one-sentence summary of Apple's latest quarter."
}
}
```
Deep research:
```json
{
"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