What problem does it solve?
PostHog Signals scouts are scheduled agents that scan a project and write reports into the Signals inbox, but understanding what the fleet is doing, why a scout stayed quiet, or whether it is healthy requires navigating configs, run rows, session logs, and scratchpad memory. This Skill provides the read-only observability playbook for answering those questions without guessing.
Core Features & Use Cases
- Fleet Surveying: List every scout's schedule, enabled status, dry-run posture, and last run, then render a scannable table with the bundled fleet_survey.py formatter.
- Run Investigation: Drill into a single run's summary, authored report IDs, and full session-log transcript, reconstructed into a readable timeline by render_run_report.py.
- Health Assessment: Judge a scout across a window of runs on cadence adherence, success rate, report rate, signal-to-noise, and memory growth using assess_health.py and the assessing-performance reference.
- Use Case: A user asks "why did my error-tracking scout find nothing this week?" — check its config posture, read recent run summaries, search the scratchpad for noise or dedupe entries, and trace whether reports were authored or silently dropped.
Quick Start
Ask the assistant to list your project's scouts and show how each one has been performing over the past few days.
