By agenkin
Claude Code plugin and skill for working with TelemetryDeck product analytics. Ask questions in plain English, get real queries, real tables, and honest caveats.
Query a TelemetryDeck app's product analytics to answer questions about DAU/MAU, retention, event counts, funnels, cohorts, and pipeline health. Use whenever the user asks a "how are users actually using this app" question that needs live data from TelemetryDeck, or is diagnosing signal-ingestion issues (missing events, opt-in anomalies, appID filters). Ships a self-contained CLI with OS-native secret storage — no repo-local `.env`, no secrets on disk outside the OS credential store.
List, switch, add, or remove the TelemetryDeck apps the CLI queries against. Accepts a subcommand as arguments (use / list / add / remove / refresh).
Daily active users for the current TelemetryDeck app. Accepts optional interval / event / compare flags as arguments.
Run the TelemetryDeck CLI end-to-end health check (platform, config, secret store, auth, query round-trip).
Schema discovery — list every event name firing for the current TelemetryDeck app with 7-day and 30-day counts side by side.
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
A Claude Code plugin (and skill) that lets agents work with TelemetryDeck product analytics. Ask in plain English, get a real query, a real table, and an honest caveat.
Once the plugin is installed, Claude knows there's an analytics skill sitting behind the /telemetrydeck-analytics:* slash commands and a CLI called tdq on PATH.
When you ask something ambient — "is the Pro share growing?", "why aren't signals arriving?", "how did last week compare to the week before?" — the analytics skill auto-invokes. Claude picks the right entry point (tdq dau, tdq groupby license, tdq query - with hand-written TQL, …), runs it, and hands you back a markdown table with the opt-in-cohort caveat attached.
When you want a specific one-shot, the namespaced slash commands do the right thing directly: /telemetrydeck-analytics:doctor for a health check, /telemetrydeck-analytics:dau --last 30d, /telemetrydeck-analytics:report "how is retention trending?", and so on.
Under the hood it's a stdlib-only Python CLI that talks to TelemetryDeck's v3 API. Secrets live in your OS-native secret store (macOS Keychain, Linux libsecret, Windows/file fallback with a loud warning) — never in a .env. A progressive-disclosure TQL reference tree means Claude reads the exact sub-page it needs instead of pasting a full spec into context.
The repo doubles as its own single-plugin marketplace. Register it, then install:
/plugin marketplace add agenkin/telemetrydeck-analytics
/plugin install telemetrydeck-analytics@telemetrydeck-analytics
Or non-interactively:
claude plugin marketplace add agenkin/telemetrydeck-analytics
claude plugin install telemetrydeck-analytics@telemetrydeck-analytics
Pulls the skill files including the bundled tdq.py — Claude uses it automatically via the $SKILL_DIR fallback. No slash commands and no tdq on PATH, but all queries work:
npx skills add agenkin/telemetrydeck-analytics
git clone https://github.com/agenkin/telemetrydeck-analytics
claude plugin marketplace add ./telemetrydeck-analytics
claude plugin install telemetrydeck-analytics@telemetrydeck-analytics
Inside a Claude Code session, run:
/telemetrydeck-analytics:setup
Claude walks you through email + password, mints a bearer, then shows a numbered picker of every app visible on your account. Pick one and you're done. Secrets land in your OS-native secret store (macOS Keychain / Linux libsecret / file fallback); nothing touches your repo.
Prefer to do it from your shell? The tdq CLI exposes the same flow — it's on PATH once the plugin is enabled:
tdq login # same interactive flow
tdq apps add <uuid> # register another app later
tdq apps use <name|uuid|index> # switch current app
setup — First run. Login, discover apps, pick the current one. Or run /telemetrydeck-analytics:setup.
doctor — Before asking real questions, verify the round-trip. Asserts non-empty rows and points at TDQ_RAW=1 on failure.
ask — Ask Claude a product-analytics question in plain English. The analytics skill auto-invokes.
narrow — When the answer needs something specific, reach for a focused command: :dau, :mau, :groupby <dim>, :events, :signals, :query.
report — For a compact markdown writeup (top-line, method, result, interpretation, caveats), use /telemetrydeck-analytics:report <question>.
iterate — If a query comes back empty or suspicious, re-run with TDQ_RAW=1 to see the raw HTTP response, or run tdq test to verify the API shape hasn't drifted.
Setup + ops
Metrics
userCount on a day granularity, not the deprecated cardinality.isAppStore, platform, custom payload keys).Schema + raw
skills/analytics/tql/index.md first.Narrative
tdq CLInpx claudepluginhub agenkin/telemetrydeck-analytics --plugin telemetrydeck-analyticsUse Amplitude like an expert - instrument analytics, discover product opportunities, analyze charts, create dashboards, manage experiments, and understand users and accounts
Access PostHog analytics, feature flags, experiments, error tracking, and insights directly from your AI coding tool. Optionally capture Claude Code sessions to PostHog LLM Analytics.
Amplitude-powered analytics skills — analyze dashboards, charts, experiments, feedback, and account health with AI.
Data analytics skills for PMs: SQL query generation and cohort analysis. Analyze user data, generate queries, and identify retention patterns.
Data & metrics skills: Data Analysis Standard, Retention Analysis, Product Health Analysis. Structure metric deep-dives, funnel analysis, cohort studies and churn investigations.
AI agent skills that make SaaS products data-ready for product analytics — from codebase scan to tracking plan to working instrumentation code.