Retrieve deduplicated signals via embedding-based HogQL queries against document_embeddings.
npx skills add https://github.com/PostHog/skills --skill signals-posthog
Please help me install this Agent Skill. Skill: signals Source: https://github.com/PostHog/skills/tree/main/skills/omnibus/signals Command: npx skills add https://github.com/PostHog/skills --skill signals-posthog
Signals are automated observations that PostHog generates by monitoring a customer's product data across multiple sources — error tracking, web analytics, experiments, session replay, and more. Each signal is a short natural-language description of something noteworthy (for example, "Error rate spiked 3× on /checkout").
Signals are grouped into Signal Reports. When a report accumulates enough weight it gets summarized and assessed for actionability. A signal report represents a cluster of related observations that together describe a meaningful issue or trend.
Signals and their embeddings are stored in the document_embeddings table, queryable via HogQL through the posthog:execute-sql MCP tool. They provide a useful way to semantically query for recent things that happened in the user's product.
The skill enables semantic and full-text querying, deduplication of signals, and targeted fetches by report or product.
The Skill's primary purpose is to help teams monitor product health by exploring signals with natural-language queries and guided SQL-like examples.
Find the top 10 signals most similar to the description 'checkout error' using semantic search.
High-intent search queries and answers about installing and using this skill.
To query PostHog signals using semantic search, you use HogQL with the embedText() and cosineDistance() functions against the document_embeddings table to rank and retrieve relevant observations.
Signals in PostHog are automated natural-language observations generated by monitoring product data. They are grouped into Signal Reports, which summarize clusters of related observations when enough weight accumulates.
To retrieve signals for a specific report, you fetch records from the document_embeddings table filtered by the target report_id, allowing deeper inspection of clustered observations.
To prevent duplicate signals, apply the argMax() function grouped by document_id. This ensures deterministic results by keeping the most recent entry based on the inserted_at timestamp.
Querying recent signals is limited to a 30-day time window. This constraint is enforced on the document_embeddings table to avoid returning stale results.
Yes, you can perform full-text keyword searches on signal content using the ILIKE operator in HogQL for fast filtering across the document_embeddings table.
PostHogQuery PostHog analytics data using HogQL and SQL foundations.
blisspixelEvaluate Primr research reports for accuracy, completeness, actionability, and citation quality.
EthGlobalBangkokIntegrate PostHog analytics into Next.js App Router with client and server instrumentation.
BEKO2210Search knowledge bases semantically with vector embeddings and keyword fallback.
