Project Planner

Project Planner

Overview

Turns a fuzzy goal into a concrete, scheduled plan. Produces a Work Breakdown Structure (WBS), milestones, an explicit dependency graph, defensible effort estimates, a critical path, and a risk register.

Keywords: project plan, work breakdown structure, WBS, milestones, dependencies, critical path, estimation, scheduling, roadmap, task decomposition, Gantt, sequencing.

Use this skill whenever the input is a goal/spec/idea and the desired output is "what needs to happen, in what order, and roughly how long".

When to use

• "Plan the migration to X" / "Break this feature into tasks" / "Build a roadmap for Q3".
• The user has a deliverable but no task list, sequence, or estimate.
• An existing task list needs dependencies, estimates, or a critical path added.

Workflow

Follow these steps in order. Do not skip clarification.

1. Clarify the goal and constraints. Capture: the single end deliverable ("done" definition), hard deadline (if any), team/resource availability, and any fixed external dependencies. Ask at most 3–5 sharp questions if these are missing. See references/clarifying-questions.md.
2. Decompose into a WBS. Break the goal top-down into deliverable-oriented work packages using the 100% rule (children fully cover the parent, no overlap). Stop decomposing at the 8/80 rule: each leaf task is 8–80 hours of work. Number hierarchically (1, 1.1, 1.1.1). See references/wbs-method.md.
3. Define milestones. Mark zero-duration checkpoints that signal a meaningful state change (e.g., "Design approved", "Beta deployed"). Every milestone must map to completed work packages.
4. Map dependencies. For each leaf task list its predecessors and the dependency type (FS/SS/FF/SF, default Finish-to-Start). Distinguish hard (mandatory) from soft (preferred) dependencies. See references/dependencies-and-critical-path.md.
5. Estimate effort. Estimate each leaf task using three-point PERT: te = (O + 4M + P) / 6. Record Optimistic/Most-likely/Pessimistic. Convert effort to duration using availability and parallelism. See references/estimation.md.
6. Compute the schedule and critical path. Forward pass for early start/finish, backward pass for late start/finish, slack = LS − ES. The critical path is the zero-slack chain; its length is the minimum project duration. Use scripts/critical_path.py to compute this automatically from your task list.
7. Identify risks. List top risks with likelihood × impact, and a mitigation or contingency buffer for each. Add schedule buffer to the critical path, not to individual tasks.
8. Produce the plan document. Fill in templates/project-plan.md. Present the WBS, milestone table, dependency/critical-path summary, estimate table, and risk register.

Estimation quick reference

• Always estimate effort (person-hours) separately from duration (calendar time). Duration = effort ÷ (people × daily-available-hours), inflated by dependencies and context-switching.
• Use PERT three-point estimates; never a single number for anything non-trivial.
• Apply a global contingency buffer (typically 15–30%) at the project level based on novelty, not per-task padding.
• Decompose anything you cannot estimate confidently — uncertainty is a signal the task is too big.

Decision heuristics

• Too big to estimate? Split it (8/80 rule).
• More than ~7 children under one parent? Add an intermediate grouping level.
• A task everyone depends on? It is likely on the critical path — protect it and start it early.
• Long pole with high uncertainty? Prototype/spike it first as its own early task to retire risk.
• Soft dependency blocking parallelism? Challenge it; reordering may shorten the critical path.

Best Practices

• Decompose by deliverables and outcomes, not by activities or org charts.
• Every leaf task has a verb, a clear "done" criterion, an owner (or role), an estimate, and predecessors.
• Make dependencies explicit; an implicit dependency is a missed delay.
• Re-estimate and recompute the critical path whenever scope changes.
• Keep the WBS and the schedule in sync; the schedule is derived from the WBS, never invented separately.

Common Pitfalls

• Confusing milestones with tasks (milestones have zero duration).
• Padding every task instead of buffering the project — hides real slack and invites Parkinson's Law.
• Listing tasks with no dependencies, producing a flat "everything at once" plan.
• Estimating duration directly and forgetting that one person can't do two parallel tasks.
• Decomposing by team ("Frontend work") rather than by deliverable, causing gaps and overlaps.
• Ignoring the critical path, then compressing non-critical tasks that don't move the deadline.

Bundled files

• references/wbs-method.md — decomposition rules, numbering, 100%/8-80 rules, worked breakdown.
• references/estimation.md — PERT, effort vs duration, story points, calibration.
• references/dependencies-and-critical-path.md — dependency types, forward/backward pass, slack, crashing/fast-tracking.
• references/clarifying-questions.md — the question checklist for step 1.
• scripts/critical_path.py — computes ES/EF/LS/LF, slack, critical path, and project duration from a JSON task list.
• templates/project-plan.md — the output document template.
• examples/website-launch.md — a full worked example from goal to scheduled plan.

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: JayRHa
• Source: JayRHa/AgentSkills
• License: MIT

Install and usage instructions live in the source repository linked above.