ID:WORK-031Status:done
plan next Command
Priority:highComplexity:moderateSource:SPEC-022
Criteria completion
No incremental history — criteria tracking started on Apr 21.
Relationships
Implements 1
Branches 3
claude/file-naming-convention-LJdwR current done
main donechangeset-release/main doneclaude/align-sidenav-styling-4MuxV done Summary
Dependency-aware work item selection. Finds the highest-priority ready item, skipping items whose dependencies aren't complete. This is the entry point of the AI agent workflow — refrakt plan next --format json tells the agent what to work on.
Acceptance Criteria
- Scans for work items with
status="ready"and bugs withstatus="confirmed" - Excludes items whose dependency IDs (from References/Dependencies section) are not yet
doneorfixed - Sorts by priority (critical > high > medium > low), then complexity (simpler items first as tiebreaker)
--milestone <name>scopes results to a specific milestone--tag <tag>,--assignee <name>,--type work|bug|allfilters work correctly--count Nreturns top N items (default: 1)--format jsonoutputs structured data including criteria list, referenced specs, dependencies, and file path- Text output shows ID, title, priority, complexity, file path, and acceptance criteria
- Exit codes: 0 = items found, 1 = no matches, 2 = invalid arguments
- Tests for filtering, sorting, dependency exclusion, and edge cases
Approach
Use the scanner (WORK-028) to get all entities. Build a status lookup map (ID → status). Filter work/bug items by ready status, then exclude those with unfinished dependencies. Sort and slice. For JSON output, include the full criteria list and spec references.
Dependencies
- WORK-027 (plugin architecture)
- WORK-028 (plan file scanner)
References
- SPEC-022 (Plan CLI)