Backlog
Unified backlog lifecycle management and orchestration across multiple storage layers (session WIP, markdown checklist files, and issue trackers).
Overview
The backlog skill provides a consistent operational interface for capturing, classifying, synchronizing, prioritizing, and maintaining backlogs across software development lifecycles.
Incoming Task / Request
├─ In-Session Short-Lived ──→ /wip (Session Task API / TodoWrite)
├─ Workspace Markdown File ──→ /fix-plan (fix_plan.md / checklist.md)
└─ Team-Shared External ──→ Issue Tracker (GitHub Issues / Forge / Secondary Trackers)
Topics
| Topic | Description | Guide |
|---|---|---|
prune |
Demote lower-priority items (P2/P3) and stale entries to preserve lean active focus sections | prune.md |
comment |
Post follow-up notes and nested review findings as comments on parent issues | comment.md |
create |
Create issues and intake items via API with fallback mechanisms | create.md |
Default Invocation (/backlog)
When invoked with no arguments, the backlog skill runs the standard health and synchronization pipeline:
- Sync: Poll external status for linked issues and pull requests to update resolved items.
- Priority Triage: Re-evaluate urgency and blockers across active items.
- Hygiene & Prune: Identify bloated or stale tasks exceeding active capacity.
- Summary Report: Output an actionable breakdown of immediate priority candidates.
Configuration
| Option | Default | Description |
|---|---|---|
--file=<path> |
fix_plan.md |
Target markdown checklist path for local operations |
--role=<pm|deep|impl> |
(unset) | Execution role profile filter for pipeline scoping |
--secondary-sync=<skill>:<topic> |
(unset) | Abstract receiver contract for non-GitHub backlog synchronization |
Cross-Workspace & Secondary Trackers
Vendor-specific and proprietary tracker integrations (e.g. Jira, Plane, internal enterprise backlogs) are decoupled from this public core skill through abstract Receiver Contracts (--secondary-sync). Private multi-repo orchestrators handle proprietary endpoints without polluting public skill schemas.
