PlanDB

Check one agent task dependency

Record two dependent tasks and see which work becomes ready when the first closes. Review the project documentation before connecting it to an agent.

What it’s useful for

PlanDB records a graph of tasks, subtasks and dependencies so an agent can find ready work and retain context between sessions. Its documented interfaces are CLI, MCP and HTTP. It belongs in a discussion of agent execution planning, rather than being treated as a screenshot-equivalent replacement for a team issue tracker.

Inside the product

• Represent nested tasks with dependencies across branches of the plan.
• Find ready work, split a task and retain discoveries or decisions for later sessions.
• Use CLI, MCP and HTTP interfaces documented by the project.

Good to know

• An agent task graph does not establish a replacement for team issue permissions, cycles, reporting or a hosted board.
• The installer can configure agent instruction files. Review that setup before installing; this review did not run it.
• No direct Linear importer or independent concurrency benchmark was verified.

The maker’s build notes

The original maker post links the repository and explicitly says PlanDB was built with Claude Code. Product features and concurrency claims come from the project’s documentation; this review did not install it, run agents or verify throughput.

A few more details

Sources & maker notes

Build details come from the maker’s linked account. vibeforthat has not independently audited the code or tested every product function.

• PlanDB — original maker development account
• PlanDB — official repository and interfaces