Discover & Doctor

Use glubean discover to see what the CLI can project from the current project, and glubean doctor to decide whether those assets and environments are ready for Cloud sync or run upload. Both commands use the same scanner; neither command syncs definitions, runs tests, or uploads evidence.

New projects created by glubean init expose the workflow as package scripts:


npm run discover   # write catalog.yml
npm run doctor     # print actionable diagnostics; non-zero when blocked
npm run sync       # glubean sync --env-file .env
npm run upload     # run the local profile and upload its evidence

Asset catalog

Running discover writes <project>/catalog.yml by default and prints a summary containing file, test, contract/case, workflow, load-plan, and OpenAPI path counts.


glubean discover
glubean discover --out catalog.json
glubean discover --out - --format yaml

The catalog includes:

recognized Glubean source files and the assets produced by each file
tests, contracts and their cases, workflows, load plans, and OpenAPI paths
projection completeness and source diagnostics
discovered .env / .env.<name> environments
project and target IDs, canonical Cloud project URLs, and separate sync and upload readiness

The catalog records tokenPresent: true|false; it never serializes the token value.

Filters

Filters are repeatable. Asset filters are combined with AND; environment filters narrow the environment list independently.


glubean discover --filter type:contract
glubean discover --filter file:'contracts/**' --filter tag:public
glubean discover --filter id:'create-*'
glubean discover --filter env:production
glubean discover --filter type:contract --filter env:production --format json

Supported selectors are type:, file:, tag:, id:, and env:. *, **, and ? wildcards are supported.

Readiness

When a token and project ID are available, discovery calls a read-only Platform API readiness endpoint. The endpoint applies the same writable-role and runs:write checks as sync/upload, resolves the upload target, and returns the canonical project URL. It does not create a projection or run.

Each environment reports one of:

Status	Meaning
`ready`	Credential, project, scope, and required target checks passed.
`blocked`	A concrete problem such as missing config, invalid token, insufficient scope, or missing target prevents the operation.
`unverified`	Local configuration exists, but exact Cloud write capability was not checked (for example, `--offline` or an older server).

Use local-only mode when network validation is unwanted:


glubean discover --offline
glubean doctor --offline

Doctor behavior

doctor prints the same inventory and readiness information but does not write a catalog unless --out is provided. Asset/import failures and concrete Cloud readiness failures exit with code 1.


npm run doctor
npm run doctor -- --filter env:production
glubean doctor --out doctor.json

discover is non-blocking by default so it can always produce the catalog. Add --strict when it should exit non-zero on the same blocking issues:


glubean discover --strict

Recommended Cloud flow


npm run doctor -- --filter env:production
npm run sync -- --env-file .env.production
npm run upload -- --env-file .env.production

Sync remains project-scoped source publication. Upload runs the generated local profile and sends target-scoped execution evidence. See Sync Projections and Upload Your First Run.