The features command manages the growth feature registry: a persistent record of all growth features detected across analysis runs.
Prerequisites
- A
feature-registry.jsonfile in the Skene bundle directory. The registry is populated by analysis runs; it must exist before exporting.
Basic usage
Export the feature registry as JSON (to stdout):
uvx skene features export
Export as Markdown to a file:
uvx skene features export --format markdown -o features.md
Export as CSV:
uvx skene features export --format csv -o features.csv
Flag reference
| Flag | Short | Description |
|---|---|---|
PATH | Project root (default: .) | |
--context PATH | -c | Path to the Skene bundle directory (skene-context/ or legacy skene/; auto-detected if omitted) |
--format TEXT | -f | Output format: json, csv, markdown (default: json) |
--output PATH | -o | Output file path. Prints to stdout if omitted. |
The feature registry
The feature registry (feature-registry.json) is automatically maintained by analysis runs. It provides:
- Persistent tracking: features are tracked across multiple analysis runs
- Merge-update semantics: new features are added, existing features are updated, missing features are archived
- Growth loop mapping: features are linked to growth loop definitions via
loop_ids - Growth pillars: features are annotated with 0-3 pillars:
onboarding,engagement,retention
Registry structure
{
"version": "1.0",
"updated_at": "2025-01-15T10:30:00",
"features": [
{
"feature_id": "team_invitations",
"feature_name": "Team Invitations",
"status": "active",
"first_seen_at": "2025-01-10T08:00:00",
"last_seen_at": "2025-01-15T10:30:00",
"loop_ids": ["invite_flow"],
"growth_pillars": ["onboarding", "engagement"]
}
],
"growth_loops": [
{
"loop_id": "invite_flow",
"name": "Invite Flow",
"linked_feature_id": "team_invitations"
}
]
}
Feature statuses
| Status | Meaning |
|---|---|
active | Feature was detected in the most recent analysis run |
archived | Feature was previously detected but not found in the latest run |
Export formats
JSON
Full registry data including all metadata, timestamps, and loop mappings.
CSV
Tabular format suitable for spreadsheets and data tools. Includes feature ID, name, status, pillars, and linked loop IDs.
Markdown
Human-readable format suitable for documentation, wikis, or README files.
Next steps
- Push: Upload the bundle (including the feature registry) to Skene Cloud
- CLI Reference: Full reference for all commands and flags
