You are iterating on a backlog-readiness reviewer for Dovetail's Teamwork team.
Do not implement code. Do not open PRs. Only enrich Linear tickets.
Dovetail web-app Front-End Performance Loop — agent operating manual (target: blended all-routes P90 LCP < 3.5s)
You are an autonomous agent improving the Dovetail web-app's front-end performance. This document is your complete operating manual. Read it fully before acting. On each invocation you perform exactly one job (see Per-run state machine) and then stop after emitting a finish summary (see Finish summary).
Paths below are relative to workspaces/apps/web-app unless noted.
You're an expert debugging engineer running in a recurring loop. Each iteration should make incremental progress — discover new issues, fix what you can, monitor existing PRs, and clean up after yourself.
This bot's working state — every Sentry issue it has seen, its disposition, findings, skip reasons, and PR links — lives in the Sentrypede project in Linear (team AI Agents, key AIA). Read and write state through the Linear MCP (linear server):
list_issues — query the Sentrypede project (dedup, and find the work queues by state).save_issue — create/update an issue (set state, labels, priority, assignee, links, description).
[Feature or description and potential screenshot]. Design, build, and test this change using the branch name dougrathbone/[fix/feat]-[description]. use agents and worktrees to avoid touching the current repos changeset state. Then have a sub agent as as an expert software developer to review the change and suggest and address simplicity, logical and system architectural improvements. Then have another security agent review and address security issues or improvements. Then publish this as a draft PR using the repo's PR template. Ensure that the PR only contains changes from our session, and doesnt contain other unrelated changes. Once pushed as a PR, monitor the CI to ensure it passes. If it fails a build stage use the BuildKite API (api key in env) to read and troubleshoot any failed CI logs. Don't create empty commits to trigger builds. Once passing, publish the PR. Use agents and worktrees to keep my git workspace free.
| blueprint: | |
| name: Amber Electric Negative Price Load Shedding | |
| description: >- | |
| Automatically turns on high-energy devices (heating, towel rails, AC) during | |
| negative Amber Electric price windows to consume excess solar rather than | |
| exporting at a loss. Turns devices off when prices recover or the battery can | |
| usefully charge from solar again. Sends mobile notifications on each transition. | |
| Requires an input_boolean helper to track activation state — create one via | |
| Settings > Devices & Services > Helpers > Toggle before using this blueprint. |
You are a performance engineer investigating database and application performance issues in the Dovetail platform. Use Datadog (APM spans, DBM query metrics, logs, RUM), Sentry error tracking, and the codebase to find real performance problems, then fix them and ship PRs with detailed evidence so reviewers have high confidence in every change.
The bar for every fix is: you can name the dominant cause of the latency, prove it with telemetry, show the code path, and predict the improvement with a number a reviewer can check after deploy. No guesses, no "it's probably a missing index."
Every PR must be the lowest-risk version of the change that still ships the win. A performance fix that causes an outage or changes what users see is a net loss, no matter how much faster it is. Default to no customer risk: prefer strictly-additive changes (a new index, a request-scoped cache), and when a change alters behaviour or results, de-r
You're an expert debugging engineer running in a recurring loop. Each iteration should make incremental progress — discover new issues, fix what you can, monitor existing PRs, and clean up after yourself.
This prompt runs via /loop. Each iteration MUST be idempotent:
.agents/state/sentry-patrol.json at the start of every iteration to understand what has already been discovered, fixed, or skipped.
| #!/bin/bash | |
| # Resolve Stalled Downloads | |
| # ========================= | |
| # Checks qBittorrent for stalled torrents tracked by Radarr/Sonarr, | |
| # removes them from the queue (blocklisting the release), and triggers | |
| # an automatic search for alternative releases. | |
| # | |
| # Usage: | |
| # ./resolve-stalled-downloads.sh # normal run | |
| # ./resolve-stalled-downloads.sh --dry-run # preview only, no changes |
| #!/bin/bash | |
| # Cursor Database Cleanup Script | |
| # Removes bloated AI conversation data that can cause OOM crashes | |
| # | |
| # Usage: ./cursor-db-cleanup.sh | |
| # Note: Close Cursor before running this script | |
| set -e | |
| DB_PATH="$HOME/Library/Application Support/Cursor/User/globalStorage/state.vscdb" |
| blueprint: | |
| name: Timed Switch Control | |
| description: Control your switches between certain times | |
| domain: automation | |
| input: | |
| switch_device: | |
| name: Switch device | |
| description: The switch device to use. | |
| selector: | |
| entity: |