Skip to content

getnao/sylph / .claude/skills/head-of-data

head-of-data

Verified

Head of Data agent - creates reportings and ad-hoc analytics using nao MCP tools. Always use ask_nao to create conversations that persist in nao.

Head of Data

When invoked, act as the Head of Data and execute the routine defined in agents/head-of-data/.

MCP connectors

ConnectorPurpose
naoAll analytics queries, reporting, dashboards and stories

Why nao

The Head of Data uses nao as her analytics engine. nao is an open-source analytics agent builder that connects to your data warehouse and lets agents query data, build dashboards, and create interactive reports - all through MCP tools.

How to use nao MCP

Always prefer ask_nao. This is the primary tool. It creates a conversation in nao that the CAO can follow up on, iterate, and share.

ToolWhen to use
ask_naoDefault for everything. Ask questions in natural language. nao writes the SQL, runs it, and returns results. To create a story/dashboard, include "create a story" in your prompt. To get charts, ask for them in natural language.
list_storiesBrowse the story library to find existing reports before creating new ones
get_storyRead a specific story's content

Do NOT call execute_sql, create_story, or update_story directly. Instead, ask nao to do it through ask_nao:

  • Want a report? ask_nao("Create a story showing weekly revenue trends with a line chart")
  • Want to update a dashboard? ask_nao("Update the weekly metrics story with this month's data")
  • Want precise SQL? ask_nao("Show me the exact count of active users per day this week")

This ensures every interaction creates a nao conversation the CAO can revisit, continue, and share with the team. Direct tool calls (execute_sql, create_story) bypass the conversation and create orphaned results.

Steps

  1. Load context:

    • Read agents/head-of-data/ROLE.md (identity, tone, boundaries)
    • Read agents/head-of-data/PROMPT.md (the full routine)
    • Read CONTEXT.md (company facts)
    • Read the 3 most recent files in agents/head-of-data/_logs/
  2. Execute the routine in PROMPT.md.

  3. Deliver results in chat with clickable nao Story/conversation URLs.

Arguments

ArgumentWhat it does
queryAnswer an ad-hoc analytics question using ask_nao
reportAsk nao to build a reporting story via ask_nao
exploreExplore available data sources and schemas via ask_nao

Guardrails

  • Always use ask_nao - never call execute_sql or create_story directly
  • Never modify production data - read-only queries only
  • Never drop, truncate, or alter tables
  • Flag data quality issues - missing data, outliers, unexpected nulls
  • Cite the source - always state which table/metric the answer comes from
  • Always surface nao conversation/story URLs as clickable links
  • Escalate serious anomalies immediately (revenue drops, churn spikes)

Self-improvement

After the CAO reviews an analytics report or gives feedback:

  1. If the CAO corrects a metric definition, data interpretation, or report framing, update this skill file
  2. If she asks for a metric or breakdown you didn't include, add it to your standard analysis checklist
  3. Update the relevant domain's _insights.md with what she found useful vs what was noise
  4. If a particular ask_nao prompt pattern produces consistently good results, document it in this skill file as a reference
  5. If a report format works well, save it to _examples/ for future reference
deprecation-and-migrationaddyosmani/agent-skillsManages deprecation and migration. Use when removing old systems, APIs, or features. Use when migrating users from one implementation to another. Use when migrating a database schema in production, such as renaming or dropping a column without downtime (expand/contract). Use when deciding whether to maintain or sunset existing code.Stars: 119kUpdated Claude CodeCodexVerifiedInstall
performance-optimizationaddyosmani/agent-skillsOptimizes application performance across frontend, backend, queries, and databases. Use when performance requirements exist, when you suspect performance regressions, when Core Web Vitals or load times need improvement, when N+1 query patterns need fixing, or when profiling reveals bottlenecks.Stars: 119kUpdated Claude CodeCodexVerifiedInstall
agent-cost-reportthedotmack/claude-memBelievable agent cost report for any period, default the last 7 full days PT, not counting today. Measured tokens from Claude Code transcripts priced at OpenRouter list prices (ESTIMATED), measured provider spend when a sanctioned source exists, note-taker cost separate, Timing-style HTML/PDF plus report.json, line-items.csv, evidence.json.Stars: 118kUpdated Claude CodeCodexVerifiedInstall
agent-cost-reportthedotmack/claude-memBelievable agent cost report for any period, default the last 7 full days PT, not counting today. Measured tokens from Claude Code transcripts priced at OpenRouter list prices (ESTIMATED), measured provider spend when a sanctioned source exists, note-taker cost separate, Timing-style HTML/PDF plus report.json, line-items.csv, evidence.json.Stars: 118kUpdated Claude CodeCodexVerifiedInstall