Task Dependency & Blocker Tracker

The problem this kills

You manage hand-offs. Task B can't start until Task A is finished, Task D waits on B and C, and somewhere in that web a single stuck task is quietly holding up six others - but nobody can see it. The status lives in a spreadsheet that goes stale the moment someone updates a cell without telling anyone. When a blocker finally clears, the person waiting downstream usually finds out days later, by accident, in a standup.

The result is invisible waiting. Work sits idle not because it's hard, but because the people who could pick it up don't know they've been unblocked. Multiply that across a project and you lose weeks to silence.

This tool makes the dependency web visible, shows you the longest chain of blocked work (your real critical path of pain), and the instant a blocker is marked done - behind a deliberate confirm step - it emails the downstream owner so they can start immediately.

What you'll build

A small, private web app for your team:

• Import your existing task list, dependency links, and people from a CSV or Google Sheet.
• A blocked-chain view that traces every task back to whatever is holding it up, and forward to everything it's holding up.
• Circular-dependency detection - it warns you when Task A waits on B, B waits on C, and C waits on A (an impossible loop), so you can fix the data instead of chasing a ghost.
• A "longest blocked chain" readout so you instantly see the deepest stack of waiting work.
• A resolve-blocker action with a human confirm step: when you mark a blocker done, the app shows you exactly which downstream tasks will unblock and who gets notified, and nothing is written or sent until you click confirm.
• Automatic downstream notifications by email (via Resend) to the owners of the newly unblocked tasks.
• A clean CSV export of updated statuses, in the exact columns your existing system expects.

What's inside the Implementation Plan

• It starts by interviewing you about your business. Before it builds anything, the plan makes the AI agent ask you about your real process - how you name tasks and IDs, where the data lives today, what "blocked" and "done" mean in your world, who owns what, your typical and peak task volumes, and your messy exceptions. It reflects a short tailored spec back to you for a thumbs-up, then shapes the data model and every later step around your answers. You get a tool fit to your team, not a generic template.
• A step-by-step build, each step ending in a ready-to-paste prompt for your AI coding agent.
• The dependency-graph logic done right: blocked-chain tracing, circular-dependency detection, and longest-chain calculation, explained in plain language.
• Duplicate guards so the same dependency pair (blocker + blocked) can't be entered twice.
• The full governance layer (below).
• A no-integration fallback so you can ship today even if you can't connect to your current system.

The governance it includes (this is the point)

This isn't a toy. The plan bakes in the controls that make an internal tool safe to actually run:

• Login so only your team can open it.
• Row-level security so each organization only ever sees its own tasks and people - enforced in the database, not just hidden in the UI.
• A complete audit trail: who resolved which blocker, who got notified, and exactly when.
• A hard human-in-the-loop gate: resolving a blocker (which unblocks others and sends emails) always requires a person to review the preview and confirm. The AI drafts the change; you approve it; only then does it commit and notify.
• Duplicate guards on the blocker/blocked dependency pair so the same link can't be processed twice.

Who it's for

Project managers, team leads, and delivery managers who coordinate hand-offs across people - and who are tired of work sitting idle because nobody knew it was ready to go. If you can use a spreadsheet, you can build and run this.

You've got this - paste the first prompt and let the agent interview you.