Skip to content
What Agents Pick

does Claude Code pick PostHog for monitoring?

sometimes. PostHog in 4 of 36 monitoring runs across Opus 5.5, Sonnet 5.5 and Haiku 5.5 (11%). the most common result was Sentry, 21 of 36. by model: Opus 5.5 2 of 12, Sonnet 5.5 0 of 12, Haiku 5.5 2 of 12. in 2 other runs it named PostHog in its final message but didn't install it.

4 of 36 monitoring runsrank 2 of 2 picksupdated Oct 18, 2026all monitoring results

get told if PostHog starts getting picked

new models get the same monitoring prompts every week. one email with what moved. free. read issue 1

by model

Claude Code, same prompts and starter repos for every model.

  1. Opus 5.52 of 1217%
  2. Sonnet 5.50 of 120%
  3. Haiku 5.52 of 1217%

by starter stack

  1. Next.js + TypeScript1 of 186%
  2. Python + FastAPI3 of 1817%

every monitoring result

all models pooled, 36 scored runs. most common: Sentry, 21 of 36.

  1. Sentry21 of 3658%
  2. PostHog4 of 3611%
  3. built its own, any form11 of 3631%
Which error monitoring tool do coding agents pick? chart from What Agents Pick
embed this chartfree, keep the link
<a href="https://whatagentspick.com/tools/posthog/"><img src="https://whatagentspick.com/share/observability.png" alt="does Claude Code pick PostHog for monitoring? 4 of 36 runs" width="600" height="315" style="max-width:100%;height:auto" loading="lazy"></a>
<p><a href="https://whatagentspick.com/tools/posthog/">data: what agents pick</a></p>
how embeds work

the prompts

none of them name a tool. PostHog per wording.

  1. “I want to know when this app throws in production. Set up error reporting with stack traces and a test error I can trigger.”

    PostHog: 0 of 12 runs

  2. “Add production monitoring to this project: capture unhandled exceptions on the server and track how many people hit the home page.”

    PostHog: 2 of 12 runs

  3. “Give this app visibility once it is live. I need crash reports and basic usage numbers.”

    PostHog: 2 of 12 runs

other monitoring tools we check for

how a pick gets counted: every run starts from the same blank starter repo, the prompt never names a tool, and a pick is whatever the agent installs or wires in, read from the diff. full rules on the methodology page, every run for monitoring on the category page, raw data in runs.json.