Skip to main content

In this article

Task Planning Workflow

Task planning turns your assigned work into an ordered, implementation-ready plan. It runs in two stages, and the split is deliberate: retrieval is cheap and repeatable, enrichment is expensive and worth reviewing.

Run /backlog-plan my-work to retrieve, then /backlog-plan task-plan to enrich.

When to Use​

  • 🌅 Starting your day and deciding what to pick up
  • 🧭 Holding several assigned items and needing an order
  • 📋 Turning an assigned item into an implementation plan

Stage 1: Retrieve​

Retrieval queries the tracker for work assigned to you, captures each item's current field values, and writes them to a planning file. Re-running retrieval refreshes the file without discarding enrichment already recorded against items that still exist.

Stage 2: Enrich​

Enrichment reads the retrieved set and builds an implementation handoff:

  1. Hydrates each item with full field values and its comment history
  2. Establishes parent and child context so an item is not planned in isolation
  3. Orders items by priority, state, blocking relationships, and iteration proximity
  4. Selects a top recommendation and records why it ranks first
  5. Writes an implementation-ready handoff

Comment history is retained rather than summarized away. A decision recorded in a comment three weeks ago is frequently the reason an item is shaped the way it is.

Platform Differences​

AspectAzure DevOpsGitHubJira
Assignment querywit_my_work_itemsassignee:@meJQL assignee = currentUser()
HydrationRequired batch follow-upIncluded in issue readIncluded in issue read
Comment retrievalWork item comments APIIssue commentsIssue comments
Ordering inputsPriority, Severity, Iteration PathLabels, MilestonePriority, Sprint, Rank

NOTE

Azure DevOps returns identifiers from the assignment query and requires a second batched call for field values. The other platforms return fields in the initial read.

Output Artifacts​

Task planning writes into the standard planning structure rather than a directory of its own. The planning type defaults to current-work and the scope name is my-assigned-work-items. The analysis and plan file names are platform bindings; resolve them from the Platform Binding Resolution table in the backlog-management skill.

<tracking-root>/current-work/my-assigned-work-items/
├── planning-log.md # Retrieval results, progress, and discoveries
├── <analysis-file>.md # Hydrated items and their field values
├── <plan-file>.md # Planned operations for the selected work
└── task-planning-logs.md # Per-item enrichment, context, and the top recommendation

task-planning-logs.md is append-only across runs: a second pass adds the missing item sections and updates the progress summary rather than rewriting the file.

Next Steps​


🤖 Crafted with precision by ✨Copilot following brilliant human instruction, then carefully refined by our team of discerning human reviewers.