Repository navigation
Closed
Labels
effort/sSmall: expected under 2 hours of workSmall: expected under 2 hours of workkind/refactorCode restructuring with no behavior changeCode restructuring with no behavior changemeta/enrichedIntake review done: acceptance criteria, links, and normalized labels addedIntake review done: acceptance criteria, links, and normalized labels addedpriority/lowNice to have; do when convenientNice to have; do when convenientsource/humanFiled directly by a humanFiled directly by a humantask/triageQueue packet kind: verify, reproduce, and route this itemQueue packet kind: verify, reproduce, and route this item
Description
Todo (Drew, 2026-11-06) — later, not now
Rename this fork's custom Claude provider from claude-code to claude-direct, aligning with the sibling custom wire-emulation providers (antigravity-direct, vibe-direct, muse-direct). The provider is our own wire emulation of the Claude Code client (built in this fork), so the -direct name is the accurate one.
Scope notes for whoever picks this up:
- The provider id is referenced by the fleet config in developing-today/code (
.opencode/opencode.jsonc— routes, pools, disabled-provider bookkeeping) and possibly by session/route-health history; the rename needs a coordinated config change in the same pass. - Consider a compat alias or a migration note for stored route_health / session rows keyed by the old id.
- No behavior change intended: wire shape, credential resolution, and quota reads stay identical.
Filed as a recorded todo per Drew's instruction; priority is deliberately low until he calls it.
Acceptance criteria
- The custom Claude provider's id is renamed per the convention decided on design: provider naming & ownership convention — mark fleet-built providers (dt- proposal), rename map #1823 (this issue's
claude-directname applies only if that convention keeps the -direct scheme), across the fork source and the fleet config in developing-today/code in the same pass. - No behavior change: wire shape, credential resolution, and quota reads are identical before and after.
- Stored route-health / session rows keyed by the old id are migrated or carry a documented compat alias.
Metadata
Metadata
Assignees
Labels
effort/sSmall: expected under 2 hours of workSmall: expected under 2 hours of workkind/refactorCode restructuring with no behavior changeCode restructuring with no behavior changemeta/enrichedIntake review done: acceptance criteria, links, and normalized labels addedIntake review done: acceptance criteria, links, and normalized labels addedpriority/lowNice to have; do when convenientNice to have; do when convenientsource/humanFiled directly by a humanFiled directly by a humantask/triageQueue packet kind: verify, reproduce, and route this itemQueue packet kind: verify, reproduce, and route this item
Activity
dezren39 commentedon Nov 6, 2026
Scope update (2026-11-06, Drew): this rename is now one line of a broader provider naming & ownership convention — see the design issue #1823 (inventory of official vs fleet-built providers, convention recommendation, full rename map, migration notes).
Under the recommended convention there,
claude-code→dt-claude(ownership prefix; the "-direct" suffix is dropped entirely because the prefix already says it's ours). This issue stays open as the claude-specific record; the rename executes from #1823 once Drew rules on the convention.dezren39 commentedon Nov 6, 2026
Intake review
🤖 Fleet monitor intake (repo-fleet-monitor), 2026-11-06.
dezren39 commentedon Nov 6, 2026
Triage (task 11114, lane opencode-a): blocked pending Drew's naming ruling.
Fresh verification 2026-11-06: #1817 is OPEN. Its own body says "later, not now" and "priority is deliberately low until he calls it", and its acceptance criteria make the
claude-directname conditional on the convention decided on #1823. #1823 is OPEN withuser/escalated+status/blocked— the provider naming/ownership convention is Drew's decision. No rename can proceed until that ruling lands; when it does, the rename map is applied as one coordinated pass across the fork source and the fleet config in developing-today/code.Queue task 11114 blocked with reason-class
user-decision.dezren39 commentedon Nov 7, 2026
Superseded by the #1823 convention as this issue's acceptance anticipated: Drew ruled (2026-11-07) for a leading underscore instead of the
-directscheme, soclaude-codeis renamed to_claude(notclaude-direct). Executed in opencode#1944 (fork source + quota) and developing-today/code#6889 (fleet config + pm-stack). Record: #1823 ruling comment #1823 (comment)dezren39 commentedon Nov 8, 2026
Triage (task 11188, lane opencode-a): already done — closing.
Fresh verification 2026-11-07 (current main 6631ec3): the rename this issue records was executed under the #1823 convention Drew ruled on 2026-11-07. The final name is (leading underscore, not ), per the superseding comment already on this issue.
Evidence:
No further rename work remains for this issue; closing as already_done / superseded by the #1823 execution.
dezren39 commentedon Nov 8, 2026
Correction to the closing note above (identifiers were stripped by shell expansion when posted): the final provider name is _claude (leading underscore, not claude-direct); at current main packages/ai/src/providers/claude-code.ts sets the provider id to _claude and packages/ai/src/quota/providers/claude.ts providerId is _claude, with claude-code retained only as an auth-source alias / wire protocol name.