refactor(cli): cover gen types runtime with effect lint (CLI-2410)

[7ttp]

TL;DR

brings the supabase gen types runtime under the effect lint

whats introduced?

effect lint applied to the gen types runtime, the tests follow separately:

• eight allow list entries in .oxlintrc.effect.json, runtime files only, so the not yet migrated test files stay out until the next branch
• the handler and the shared helpers yield their tagged errors directly instead of wrapping them in Effect.fail, every message left byte for byte the same
• the flag rejections, the local db inspect failures and the missing preview branch credentials become tagged errors, which adds four names to the error tag fixture and makes --output-format json report their tag as error.code instead of UnknownError
• an unreachable container runtime is tagged apart from a stopped stack in telemetry, rather than both landing in the same bucket, and the terminal text is unchanged either way
• the generator layer splits its steps so interruption lands between them, and the duration quote encodes through Schema

ref:

• related: CLI-2410

[github-actions]

🤖 AI Review

Both findings are confirmed. The exit-code table omits the positional --lang rejection and does not name the new error tags for several listed conditions. Unrelated Docker inspection failures receive telemetry suggesting supabase start. No additional issues surfaced.

Findings

Severity	Location	Category	Sources	Claim
🟡 MINOR	apps/cli/src/commands/gen/types/types.handler.ts:508	telemetry-classification	codex	An inspection failure unrelated to container state is classified in telemetry as a local stack that needs supabase start.
⚪ NIT	apps/cli/src/commands/gen/types/SIDE_EFFECTS.md:87	documentation	claude	The exit-code table leaves the new flag-usage and local-database error tags unnamed, and has no row for the positional --lang rejection.

Findings outside the diff

• ⚪ NIT apps/cli/src/commands/gen/types/SIDE_EFFECTS.md:87 — The exit-code table leaves the new flag-usage and local-database error tags unnamed, and has no row for the positional --lang rejection.

Stats

Claude findings: 1 · Codex findings: 1 · Confirmed: 2 · Refuted: 0 · Uncertain: 0

---

Models: claude-opus-5-5 + gpt-6-sol · Trigger: auto · Workflow run

This review runs once per PR. A maintainer can request another with a /ai-review comment.

[Coly010]

This is a nice step up - -o json reporting real tags instead of UnknownError is a genuine improvement, and I checked every converted message against base and they're byte-identical ("must used" typo included). The 743b698 fix looks right to me as well: "no such container" still gets the supabase start message, everything else lands on the inspect error without it

Approving, with two asks. Both are about the new error.code values, since they become something scripts can branch on once this ships:

1. inline, on the preview branch credentials tag
2. SIDE_EFFECTS.md - can't comment inline as it's not in the diff, but the exit code table (L84-96) still doesn't have rows for the positional --lang rejection or missing preview branch credentials, and none of the new tags (GenTypesFlagUsageError, GenTypesLocalDbNotRunningError, GenTypesLocalDbInspectError) are named. The "not running or db inspection failed" row also now covers two tags that behave differently. I think that table is the only place the UnknownError → tag change would get recorded, so it'd be good to land it with this PR