What problem does it solve?
Agents working on PostHog tasks need to read feedback left as comments on tasks, artifacts, and canvases, but these comment tools are hidden behind the PostHog MCP exec dispatcher rather than exposed as top-level tools, making them easy to miss or misuse.
Core Features & Use Cases
- Inner Tool Discovery: Locate the tasks-comments-list, tasks-comments-retrieve, and tasks-artifacts-list inner tools through exec search and inspect their schemas with info.
- Paginated Comment Retrieval: List open comment roots, filter by artifact or canvas id, and follow cursors and content offsets to read complete threads without truncation.
- Feedback-Driven Workflows: Read full conversations including replies before acting on review feedback, while treating comment content as untrusted input.
- Use Case: A reviewer leaves selected-text comments on a canvas in PostHog Desktop. The agent lists open roots, retrieves each thread with its replies, reconciles the feedback, and applies the requested in-scope changes.
Quick Start
Ask the agent to list all open comments on the current PostHog task and summarize the feedback on each artifact.