القائمة
Query Honeycomb for SLO compliance, error budgets, burn alerts, and OpenTelemetry traces. Use when asked about SLO status, service health, error budgets, latency, or Honeycomb data.
التثبيت باستخدام Codex أو Claude انسخ هذا Prompt والصقه في Codex أو Claude أو مساعد آخر ليراجع صفحة Skill ويثبّتها لك.
استنادا إلى تصنيف SOC المهني
Fetch free long-term U.S. company financial data from the SEC EDGAR REST API — filing histories, company metadata (CIK, SIC, tickers, exchanges), and standardized XBRL financials (revenue, assets, EPS, cash flow) over many years for any U.S. filer. Use this skill when the user asks about a public company's SEC filings, fundamentals, historical financial statements, or wants to compare a metric across companies, even if they don't mention EDGAR, the SEC, or XBRL by name. No API key required.
Fetch Euronext stock option chains — search for optionable instruments, discover available expiry dates, and pull full option chains (calls/puts with strikes, bids, asks, lasts, settlements) for any Euronext-listed equity option. Use this skill when the user asks about European stock options, Euronext option chains, LEAPS on European equities, or wants to scan option prices across expiries for stocks like Shell, TotalEnergies, Airbus, or Heineken, even if they don't mention Euronext by name.
Query ACARA My School data for every Australian school — school profiles, student-to-teacher ratios, neighbourhood demographic indicators (ICSEA), enrolments, and locations with coordinates. Use this skill when the user asks about schools near a property, what schools exist in a postcode or suburb, student-to-teacher ratios, or neighbourhood character via school demographics, even if they don't mention ACARA or My School directly.
Fetch Japanese stock option chain data from the Osaka Exchange via JPX. Use this skill when the user wants to see option prices, implied volatility, or strikes for a listed Japanese company — including evaluating protective puts, covered calls, or hedging strategies for TSE stocks. Accepts a 4-digit security code (e.g., 6146 for Disco Corp, 7203 for Toyota). Supports both the nearest and subsequent monthly expiries.
Create detailed business overviews for investment research — deep-dive through business model, management, financials, valuation, risks and tracking. Use this skill only when the user explicitly asks to do a business overview, deep dive, or formal company write-up.
Evaluate spinoffs, merger securities, post-bankruptcy orphan equities, corporate restructurings, stub stocks, recapitalizations, rights offerings, and LEAPS/warrants in extraordinary corporate events. Use this skill when the user asks about a company undergoing a corporate change, mentions a spinoff, merger consideration, post-reorg stock, going-private transaction, Dutch auction tender, rights offering, or corporate breakup, even if they don't name the event type explicitly.
| name | read-honeycomb-opentelemetry |
| description | Query Honeycomb for SLO compliance, error budgets, burn alerts, and OpenTelemetry traces. Use when asked about SLO status, service health, error budgets, latency, or Honeycomb data. |
| allowed-tools | mcp__honeycomb__get_slos, mcp__honeycomb__get_triggers, mcp__honeycomb__run_query, mcp__honeycomb__run_bubbleup, mcp__honeycomb__get_trace, mcp__honeycomb__find_columns, mcp__honeycomb__get_workspace_context, mcp__honeycomb__get_environment, mcp__honeycomb__get_dataset |
READ-ONLY observability skill. Query SLO status, error budgets, burn alerts, and traces.
📖 Full guide: Remote's SLOs framework (Notion)
If no Honeycomb MCP is configured:
claude mcp add honeycomb --transport http https://mcp.eu1.honeycomb.io/mcp
Only READ access is needed during OAuth setup.
Always start by discovering what's available:
get_workspace_context → list environments and datasets.get_environment → confirm the target environment (typically production).get_dataset → confirm the target dataset.Skip 2–3 if the user specifies environment and dataset explicitly.
Use get_slos with environment_slug to list all SLOs.
Triage pattern:
run_query to investigate contributing errors or latency.Use get_triggers with environment_slug to list burn alert triggers and current state.
Use run_query with environment_slug and dataset_slug.
Error rates: COUNT where error = true or status_code >= 500, broken down by service.name or endpoint. Use RATE_AVG or COUNT with error filter vs total for error percentage.
Latency: HEATMAP(duration_ms) or P99(duration_ms) / P95(duration_ms) / P50(duration_ms). Break down by endpoint or operation.
Availability: Total requests vs error requests over a time window.
Time windowing: For incidents, narrow to the incident window first, then expand if needed.
Use run_bubbleup to identify what makes a problematic subset different from the baseline.
When to use: SLO breaching, post-deploy comparison, latency spikes.
Pattern:
duration_ms > P99 or error = true).MCP responses include URLs (trace_link from get_trace, Query Url from run_query). Always prefer those over manually constructing URLs.
15-second frontend timeout: Dragon cuts requests at 15s. Traces for long requests may appear incomplete, child spans disconnect from the parent. Gaps in span hierarchy on 15s+ traces are likely timeout-related, not missing instrumentation.
production environment unless told otherwise. Discover the dataset via bootstrap.