01 Reference
CLI overview
A bounded command reference for installation, offline learning and signed evidence.
Install tools through a reviewed plan
These commands come from the documented library and Analytical Spine workflows. Run library commands from an approved checkout. Run installed ledger commands from the target project with the required retained proof. Angle-bracket arguments represent reviewed values; the examples are not ready-to-run project identities.
node tools/agent-skills.mjs validate suites/analytics-spine.jsonnode tools/agent-skills.mjs plan suites/analytics-spine.json --target <target-repository> --profile full --agents openai-codexValidate checks the suite contract. Plan describes target operations, selected adapters and required parameters. After approving the writes, use the matching apply command and then verify the installation. Readiness reports the installed capability state; it does not qualify a pipeline.
node tools/agent-skills.mjs apply suites/analytics-spine.json --target <target-repository> --profile full --agents openai-codexnode tools/agent-skills.mjs verify suites/analytics-spine.json --target <target-repository> --profile full --agents openai-codexnode tools/agent-skills.mjs readiness --target <target-repository>Use guided skills for the engineering steps
The installed front door is /analytics-spine. In Claude Code, /analytics-spine:analytics-spine is the namespaced alternative. Skills run when explicitly invoked. Each operation still needs the business facts, current evidence and permissions appropriate to its boundary.
| Skill | Documented purpose |
|---|---|
| /source-map | Profile or extract under the selected source-read workflow. |
| /source-oracle | Capture product actuals for comparison with independent expectations. |
| /domain-index and /oltp-to-star | Review process meaning and draft a star contract for human grain ratification. |
| /medallion and /fabric-run | Compile and execute a reviewed, scoped Fabric plan. |
| /semantic-model and /dax-measure | Generate TMDL and probe measures before committing them. |
| /pbi-page and /pbi-deploy | Author PBIR and review exact deployment plans. |
| /data-atlas | Render and inspect local lineage, evidence and qualification state. |
This table is a map, not an unattended script. Deployment, source access and model acceptance must not be inferred from the order of the rows. Use the owning workflow's current instructions for each effect.
Learn without a tenant
node packages/powerbi/dimensional-model/tests/demo.mjsnode packages/data/data-ledger/tests/regressions.mjsThe first command demonstrates fixture modelling and semantic metadata generation. The second includes an intentionally all-red missing-evidence gate. The regression passing means the refusal worked; it does not mean a live pipeline passed. Read Your first qualification for the expected diagnostic and interpretation.
Authenticate the retained evidence and projection
For an existing signed proof with independently supplied public trust, strict verification checks the retained bindings. Carry the original context for scoped proof; do not replace it with a convenient new scope.
node .agent-skills/data/ledger/spine.mjs verify --project <project> --evidence --strict --require-signatures --trust-store <public-trust-file>An exported value-free attestation can be checked independently:
node .agent-skills/data/ledger/spine.mjs verify-attestation --attestation <new-attestation-file> --trust-store <public-trust-file>That authenticates the envelope under the selected trust. It does not reveal the private values or recompute their business comparison. Export and push have additional subject, custody, expiry and authorisation prerequisites; do not reduce them to a generic upload command.
Keep permission next to the operation
The general Agent Spine setup workflow serves repository orchestration. It is distinct from Analytical Spine's analytics qualification front door. Choose the command for the actual task and consult the limits guide before describing its result as proof.