Overview
Thepipeline-state.json file tracks the execution status of all stages in the SDD pipeline. It is created by the setup skill and automatically updated by the state-updater hook (H3) after each skill execution.
Location
Project root:pipeline-state.json
Schema version
Current: v2.0 (introduced in SDD Plugin v2.0.0)Root fields
The active pipeline stage. Typically the next stage that should be executed.Valid values:
"requirements-engineer""specifications-engineer""spec-auditor""test-planner""plan-architect""task-generator""task-implementer"
ISO-8601 timestamp of the last pipeline state modification.Example:
"2026-03-01T15:30:00.000Z"Map of stage names to their execution status. Each key is a stage identifier.
Example
Usage
Initialization
Created by/sdd:setup with all stages set to "pending":
Automatic updates
Thestate-updater hook (H3) automatically updates the pipeline state after each skill execution:
- Sets the completed stage to
"done" - Updates
outputHashbased on generated artifacts - Advances
currentStageto the next pipeline stage - Updates
lastUpdatedtimestamp
Manual queries
Check current pipeline status:Constraints
- One pipeline per project: Only one
pipeline-state.jsonshould exist in the project root - Never delete manually: The pipeline state is authoritative for upstream guards (H2)
- Hash stability: The
outputHashis generated from file contents, not timestamps - Staleness detection: Stages become
"stale"when their input artifacts change
Related
- Skills:
/sdd:setup,/sdd:pipeline-status - Hooks: H3 (state-updater), H2 (upstream-guard)
- MCP resources:
sdd://pipeline/status,sdd://pipeline/stages