froggo-mission-control: Skill for Claude Code

Data Analysis

Purpose

Produce analyses that answer specific questions and inform specific decisions. Every analysis starts with a clear question and ends with a recommendation — not just observations. Data without a "so what" is noise.

Trigger Conditions

Load this skill when:

• A request for data analysis, a report, or a metrics deep-dive arrives
• Investigating a performance anomaly or unexpected metric change
• Validating whether a product change or campaign produced the expected result
• Producing a periodic report (weekly, monthly, quarterly)
• Answering a specific business question with quantitative data

Procedure

Step 1 — Write the Analysis Brief

Before touching any data, answer these questions in writing. If you cannot answer them, the request is not ready to be worked on — route back to the requester.

Analysis brief:

Question being answered:
[One precise question. "How is the app performing?" is not a question. "What is the 30-day trend in D7 retention rate for users who completed KYC in January?" is a question.]

Decision this informs:
[What will someone do differently based on the answer? If the answer is "nothing" or "it depends," the question may be the wrong question.]

Requester:
[Who asked? Agent or human name]

Deadline:
[When is the answer needed?]

Acceptable data sources:
[Internal DB? Mixpanel? Third-party API? Scraped data? Internal-only or public data OK?]

Output format expected:
[Summary report? CSV export? Chart? Dashboard update? Inline task activity?]

Step 2 — Identify Data Sources

List every source needed before writing a single query. For each:

Source 1:
  - Name / tool: [e.g., mission-control DB, Mixpanel, on-chain data API]
  - Access method: [MCP tool / SQL / API endpoint]
  - Tables or events relevant: [e.g., `users`, `transactions`, `funnel_events`]
  - Date range available: [how far back does data go?]
  - Known limitations: [e.g., "only tracks web, not mobile," "30-day lookback limit"]

Source 2: ...