selftune team

selftune team is the noninteractive CLI for the same Skill Set releases, assignments, contributions, and lifecycle state shown in SelfTune Desktop and Cloud. It does not use the legacy compatible Registry state model. Set the hosted endpoint and provide credentials without putting them in shell arguments:

export SELFTUNE_CLOUD_URL="https://your-deployment.convex.site"
export SELFTUNE_SERVICE_TOKEN="..."
selftune team status --release-id release_123 --json

For a one-shot secret, pipe it to --token-stdin. Contribution submission uses the linked Desktop credential from SELFTUNE_DEVICE_TOKEN or stdin because the server must bind the proposal to the exact recipient assignment. Tokens are never accepted as command arguments or printed.

Commands

selftune team publish --skill-set-id support --envelope release.json \
  --revision-sha256 <sha256> --envelope-sha256 <sha256> --yes --json
selftune team assign --request-id assign-123 --release-id release_123 \
  --member-id member_123 --device-id device_123 --update-policy notify --yes --json
selftune team status --release-id release_123 --json
selftune team contribute --request-id proposal-123 --skill-set-id support \
  --base-release-id release_123 --envelope proposal.json \
  --revision-sha256 <sha256> --envelope-sha256 <sha256> --title "Fix" --yes --json
selftune team promote --release-id release_123 --revision-sha256 <reviewed-sha256> \
  --envelope-sha256 <reviewed-sha256> --yes --json
selftune team deprecate --release-id release_123 --reason "Superseded" --yes --json
selftune team rollback --request-id rollback-123 --release-id release_122 \
  --member-id member_123 --device-id device_123 --reason "Regression" --yes --json

All state-changing commands refuse to run without --yes. Exit code 0 means success, 1 an operational failure, 2 invalid/missing arguments or missing confirmation, and 3 rejected credentials. --json emits one JSON value to standard output; diagnostics go to standard error.

Start

Manage your skills

Improve and validate

Run it your way

For agents

Concepts

Operate safely

Reference

Commands

​Commands

Commands