posthog

Explore and diagnose a PostHog endpoint's execution logs — error messages, failed runs, cache misses, slow runs, or unexpected row counts during endpoint invocations. Use when the user says "my endpoint is failing", "show me the logs for endpoint X", "what error did endpoint Y produce", "why did endpoint Z return no rows", "is this endpoint hitting cache", or "check the last N runs". Focused on a single named endpoint's runtime log entries, not project-wide auditing or query performance profiling.

Diagnose why a PostHog endpoint is slow or expensive and propose a concrete fix — bump the cache TTL, enable materialisation, restructure variables, or rewrite the query. Use when the user says "this endpoint is slow", "my endpoint times out", "we're hitting the cost cap on this one", or asks "should I materialise this?". Focuses on a single named endpoint, not a project-wide audit.

Create a PostHog endpoint with the right shape on the first try — covers query kind choice, name conventions, what to expose as variables (HogQL code_name vs insight breakdown), data_freshness_seconds, and whether to materialise on day one. Use when the user says "create an endpoint", "expose this query as an API", "turn this insight into an endpoint", or asks for help structuring a new endpoint. Steers away from common mistakes: materialising a query with cohort breakdowns or compare mode, inline-only variables on a materialised endpoint, unbounded date ranges, ambiguous names.

Copy a PostHog endpoint (a saved HogQL/insight query exposed as an API route) to another project in the same organization, or duplicate it under a new name in the same project. Use when the user wants to duplicate an endpoint, promote an endpoint from staging to production, replicate an endpoint's query/variables/freshness config in another workspace, or clone an endpoint to iterate on it. Unlike feature flags and experiments, endpoints have NO native cross-project copy tool — this skill covers the read-then-recreate flow (endpoint-get then endpoint-create), the active-project switching it requires, name-collision checks, and the safe defaults (land unmaterialised in the target, verify with endpoint-run). Does not cover editing endpoint versions (see managing-endpoint-versions) or authoring a brand-new endpoint from scratch (see creating-an-endpoint).

Wire a PostHog endpoint into a client app or SDK. Covers fetching the OpenAPI spec, generating a typed client with openapi-generator or @hey-api/openapi-ts, sending the right auth header, shaping the variables payload (HogQL code_name vs insight breakdown property), handling rate-limit and materialised-endpoint error responses. Use when the user says "how do I call my endpoint", "generate a client for this", or "what auth header do I use".

Audit every endpoint in a PostHog project for staleness, failed materialisations, and unused materialised versions. Use when the user asks "what endpoints can I clean up?", "are any of my endpoints broken?", "which materialised versions are still being called?", or wants a one-shot cleanup pass over the Endpoints product. Produces a prioritised report grouped by issue type, with recommended actions but does not modify anything without explicit confirmation.

Audit the health of a PostHog project's materialized views (saved queries) — find every failed materialization and flag unused or stale materialized views that cost storage and compute. Use when the user asks "which of my views are broken?", "why is this materialized view failing?", "are any of my views wasting compute?", or wants a one-shot triage of view health. For source/sync health use `auditing-warehouse-source-health`.

Build a new dashboard, or update an existing one, from a set of insights — the same job the in-app assistant does with its upsert-dashboard tool, but over MCP. Use when a user asks to create a dashboard, put several metrics/charts together on one page, assemble a dashboard for a topic (product analytics, retention, revenue, activation, etc.), or add/remove/replace insights on a dashboard they already have. Covers deciding create vs update, reusing existing insights vs creating new ones, and using PostHog's vetted dashboard templates as reference for what a strong dashboard on a topic looks like.

Export PostHog events, persons, or sessions on demand and download the resulting files. Use when the user asks to download/export raw PostHog data, create a one-off file export, fetch a Parquet or JSONLines export, or use the file_download_batch_exports API. Covers starting the export with MCP, polling completion, and downloading via the existing REST redirect endpoint.

Set up an LLM-judge evaluation that extracts canonical use cases for a PostHog feature at scale and streams the results to a Slack channel as a live feed. Use when someone wants to understand how users are actually using a specific AI/LLM-powered feature in production — what they're investigating, what questions they're trying to answer, and what patterns surface — without manually reading hundreds of traces. Assumes the feature emits `$ai_generation` and `$ai_evaluation` events with `$session_id` linkage to the trigger user's recording (the standard setup post the session-summary linkage PRs).

ABSOLUTE MUST to debug and inspect LLM/AI agent traces using PostHog's MCP tools. Use when the user pastes a trace or session URL (e.g. /ai-observability/traces/<id> or /ai-observability/sessions/<id>), asks to debug a trace, figure out what went wrong, check if an agent used a tool correctly, verify context/files were surfaced, inspect subagent behavior, investigate LLM decisions, or analyze token usage and costs. Also use when raw SQL/HogQL against `events.properties.$ai_input` / `$ai_output_choices` returns empty — message content lives only on the dedicated `posthog.ai_events` table.

Investigate AI observability evaluations — `hog` (deterministic code-based), `llm_judge` (LLM-prompt-based), and `sentiment` (user-message sentiment). Find existing evaluations, inspect their configuration, run them against specific generations, query individual results, and generate AI-powered summaries for boolean pass/fail runs. Use when the user asks to debug why an evaluation is failing, surface common failure modes, compare results across filters, dry-run a Hog evaluator, prototype a new LLM-judge prompt, inspect sentiment classifications, or manage the evaluation lifecycle.

Investigate LLM spend in PostHog — total cost over time, cost by model, provider, user, trace, or custom dimension, token and cache-hit economics, and cost regressions. Use when the user asks "how much are we spending on LLMs?", "which model / user / feature is most expensive?", "why did cost spike?", wants to build a cost dashboard or alert, or pastes a trace URL and asks about its cost.

Investigate AI observability clusters — understand usage patterns in AI/LLM traffic, compare cluster behavior, compute cost/latency metrics, and drill into individual traces within clusters.

Find where an AI/LLM application is failing in production and surface the failure patterns, working from real traces. Use when someone wants to understand what's going wrong with an AI feature, find and categorize failure modes, triage errors, or investigate quality issues (wrong answers, ignored instructions, hallucinations, tool misuse) — "what's failing in my agent", "surface error patterns", "why are the responses bad", "find the common failure modes", "what should I fix next". Covers scoping to one use case, finding failing traces by whichever signal fits the context (code errors, metric outliers, trace-type slices, manual review, existing-eval spikes, clustering), and reading them into a ranked failure taxonomy.

Author continuously-running online evaluations in PostHog AI observability, grounded in a real failure mode you've identified. Use when the user wants an evaluation that automatically scores new generations or whole traces going forward — "create an eval to catch X", "continuously check that responses do Y", "turn this failure into an eval". Covers choosing the target and eval type (hog / llm_judge / sentiment), gating on the team's provider key before an llm_judge eval, scoping which generations trigger it via conditions (property filters + rollout sampling), creating it disabled, verifying scope, and enabling. Finding and ranking the failure modes worth evaluating is its own job — use exploring-ai-failures first. To debug or manage evaluations that already exist, use exploring-llm-evaluations.

Analyze the most expensive users in AI observability and explain why they cost so much. Use when the user asks about top spenders, expensive users, per-user LLM cost, user-level cost drivers, or patterns behind high AI observability spend.

The fleet-wide audit — sweep every agent, mine recent sessions for failures/anomalies, classify root causes, and branch a DRAFT proposal per fix (never freeze/promote); write a report to memory. Load when asked to audit all agents or what's underperforming.

How to instrument PostHog's own Metrics product from PostHog-owned code — record counters, gauges, and histograms that land in posthog.metrics, the same way customers do. Use when adding application metrics in this monorepo (web, Celery, Temporal), when asked to push or ship metrics into posthog metrics, or when unsure whether the SDK in this environment supports posthog.metrics yet. Covers the environment decision (SDK-first per the public docs, OTel fallback when the SDK path is not available), the exact version gates per SDK, what is already wired internally, and how to validate metrics actually arrive.

Ingestion pipeline architecture overview and convention reference. Use when you need a quick orientation to the pipeline framework or want to know which doctor agent to use for a specific concern.

Use when editing, reviewing, or auditing DRF viewsets and serializers in PostHog. Triggers on files in posthog/api/, products/*/backend/api/, products/*/backend/presentation/, or any file importing rest_framework. Covers field typing, schema annotations, enum collision fixes, and OpenAPI spec quality — everything that flows downstream into generated TypeScript types and MCP tools.

Implement and extend PostHog Data warehouse import sources. Use when adding a new source under products/warehouse_sources/backend/temporal/data_imports/sources, adding datasets/endpoints to an existing source, or adding incremental sync, resumable imports, webhook ingestion, pagination, credentials validation, and source tests.

Guide for adding MCP UI apps — interactive visualizations that render tool results in MCP clients like Claude Desktop. Use when adding a new detail or list view for an MCP tool, creating view components in products/*/mcp/apps/, or linking tools to UI apps via YAML.

Guide for exposing PostHog product endpoints as MCP tools. Use when creating new or updating API endpoints, adding MCP tool definitions, scaffolding YAML configs, or writing serializers with good descriptions. Covers the full pipeline from Django serializer to generated TypeScript tool handler.

Guidelines to create/update a new mode for PostHog AI agent. Modes are a way to limit what tools, prompts, and prompt injections are applied and under what conditions. Achieve better results using your plan mode.