| name | qlik-project-scaffold |
| description | Initialize a new Qlik project with standard directory structure, input subdirectories, artifact phases, and dependency tracking template. |
| disable-model-invocation | true |
qlik-project-scaffold
Overview
This skill scaffolds a new Qlik development project with the complete directory structure specified by the Qlik Development Framework. It creates the inputs/ directory tree for source materials, the artifacts/ directory tree for framework output organized by phase, and initializes .pipeline-state.json for dependency tracking throughout the pipeline.
The skill is procedural (no model invocation) and idempotent—safe to re-run on an existing project without destroying data.
What this skill creates
1. Input Source Directories
inputs/source-documentation/ — Connection specifications, upstream data dictionaries, platform documentation
inputs/existing-apps/ — QVF files, load scripts, metadata from brownfield applications
inputs/platform-libraries/ — Shared Qlik libraries, naming conventions, reference implementations
inputs/upstream-architecture/ — ER diagrams, data lineage, ETL architecture from source systems
Each subdirectory includes a README.md explaining what materials belong there.
2. Artifact Directories by Phase
artifacts/00-platform-context.md — Phase 0 output (framework context document)
artifacts/01-project-specification.md — Phase 1 output
artifacts/02-source-profile.md — Phase 2 output
artifacts/03-data-model-specification.md — Phase 3 output
artifacts/04-scripts/ — Phase 4 outputs (load scripts, manifest, diagnostics/)
artifacts/05-expression-catalog.md — Phase 5 output
artifacts/05-expression-variables.qvs — Phase 5 expression variables
artifacts/06-viz-specifications.md — Phase 6 outputs
artifacts/07-qa-reports/ — Phase 7 comprehensive QA review
artifacts/08-documentation/ — Phase 8 documentation outputs
3. Dependency Tracking
.pipeline-state.json — JSON state file tracking current phase, artifact versions, blocked dependencies, and execution validation requests/results
Usage
From your Qlik project root directory:
/qlik-project-scaffold
The skill creates all directories and the initial .pipeline-state.json template. Existing files are not overwritten.
Directory Structure Created
project-root/
├── inputs/
│ ├── source-documentation/
│ │ └── README.md
│ ├── existing-apps/
│ │ └── README.md
│ ├── platform-libraries/
│ │ └── README.md
│ └── upstream-architecture/
│ └── README.md
├── artifacts/
│ ├── 04-scripts/
│ │ └── diagnostics/
│ ├── 07-qa-reports/
│ └── 08-documentation/
└── .pipeline-state.json
Output
Upon completion:
✓ Qlik project structure initialized
✓ Input directories created (source-documentation, existing-apps,
platform-libraries, upstream-architecture)
✓ Artifact directories created (04-scripts/diagnostics, 07-qa-reports,
08-documentation)
✓ Pipeline state file created: .pipeline-state.json
Next steps:
1. Place source materials in inputs/ subdirectories
2. Review .pipeline-state.json
3. Start Phase 0 with requirements-analyst
Procedure
Step 1: Create input directories
mkdir -p inputs/source-documentation
mkdir -p inputs/existing-apps
mkdir -p inputs/platform-libraries
mkdir -p inputs/upstream-architecture
Step 2: Create artifact directories
mkdir -p artifacts/04-scripts/diagnostics
mkdir -p artifacts/07-qa-reports
mkdir -p artifacts/08-documentation
Step 3: Create input README templates
Each inputs/ subdirectory receives a README.md describing what materials belong there.
Step 4: Initialize .pipeline-state.json
Create the dependency tracking template:
{
"currentPhase": 0,
"projectName": "qlik-project",
"projectDescription": "",
"mcpAvailable": {
"qlikCloud": null
},
"artifacts": {},
"blockedDependencies": [],
"executionValidation": {
"requests": [],
"results": []
},
"notes": "Initialize with Phase 0 context analysis. Update currentPhase and artifacts after each phase completion. Set mcpAvailable.qlikCloud to true/false at session startup."
}
Notes
- Idempotent: Safe to run multiple times. Existing files are preserved.
- Plugin-supplied: Agents and skills are provided by the qlik-agents plugin. This skill only creates the project-specific directory structure.
- Read-only inputs: All materials in
inputs/ are read-only for agents during pipeline execution.
- Artifact evolution: Agents write to
artifacts/ only. The orchestrator coordinates phase transitions and artifact updates.
- Field naming: All field examples in scaffolded documentation follow entity-prefix dot notation (e.g.,
[Customer.Status], [Order.Amount]).