Repository navigation
Decide which plugin features should ship as mods #5778
Copy link
Copy link
Closed
Labels
needs-humanHuman-in-the-loop required; autonomous sessions must not resolve items carrying this.Human-in-the-loop required; autonomous sessions must not resolve items carrying this.priority: lowNice-to-have, cosmetic, or speculative; opportunistic.Nice-to-have, cosmetic, or speculative; opportunistic.status: needs-decisionAwaiting a human or maintainer judgment call.Awaiting a human or maintainer judgment call.work-class: structuralRefactors, migrations, contract changes; cross-cutting and hard to reverse.Refactors, migrations, contract changes; cross-cutting and hard to reverse.
Description
Goal
Decide which of our plugin features would be better shipped as a mod than as a skill or a settings hook, and how the marketplace should carry mods.
Context
- Since Claude Code 2.1.287 a marketplace plugin can ship a mod: function hooks that hot-reload and can draw panes, bands, status lines and toasts, and can hook events such as
turn.step,prompt.submitandsession.measure. - Claude Code itself ships 11 built-in mods (
<name>@builtin), e.g.cc-plugin-agents-md,cc-plugin-diff,cc-plugin-you-should-know. - The API is early access and may change between releases without notice. The built-in
plugin-authoringskill is regenerated per build with that build's types, so we point at it rather than copying it. - A 2026-10-01 overlap review recommended keeping desktop-notification (mod toasts only show inside Claude Code; mod audio is silent on Linux and Windows), guardrails, the formatters, the session-flow observer and the user status line as they are, and piloting context-guard/rate-limit-guard first (separate issue).
Questions to settle
- Which plugins have hook-based behaviour that a mod would do better (lower process fan-out, in-UI surfaces, coverage in
-p/cloud sessions)? - How do we version and test mods given the early-access API (a release-recheck cadence,
claude plugin testin CI)? - Do mods change how we split skills vs hooks in the plugin philosophy doc?
Start after the context-guard pilot reports what worked.
Reactions are currently unavailable
Metadata
Metadata
Assignees
Labels
needs-humanHuman-in-the-loop required; autonomous sessions must not resolve items carrying this.Human-in-the-loop required; autonomous sessions must not resolve items carrying this.priority: lowNice-to-have, cosmetic, or speculative; opportunistic.Nice-to-have, cosmetic, or speculative; opportunistic.status: needs-decisionAwaiting a human or maintainer judgment call.Awaiting a human or maintainer judgment call.work-class: structuralRefactors, migrations, contract changes; cross-cutting and hard to reverse.Refactors, migrations, contract changes; cross-cutting and hard to reverse.
Activity
kyle-sexton commentedon Oct 4, 2026
This was generated by AI during triage.
Routing: human-gated,
status: needs-decision, work class structural (C4). This is a product and architecture decision (which plugin features move to a new runtime surface, how the marketplace versions and tests it, and whether the skills-versus-hooks split in the plugin philosophy changes), and any adoption supersedes ADR 0035's Defer.Basis: docs/adr/0035-defer-claude-code-mods-with-five-go-criteria.md (Defer; no plugin gains a
moduleskey until all five go criteria hold); docs/upstream/claude-code-mods/go-no-go.md ("After a run": a flip to go needs a superseding record and reopens the guard-conversion conditions rather than lifting them).Sequencing, by the issue's own text: start after the context-guard pilot (#5777) reports. Upstream blocker today: go criterion 3 fails (anthropics/claude-code#92533 is OPEN, checked 2026-10-04), so no answer here can ship yet. #5768 re-runs the go/no-go runbook. No blocked-by edges were added; the chain is reported to the maintainer.
Already in the repository, for when this is picked up: the mod-authoring convention (how a marketplace plugin would ship a mod: version floor, mod versus settings hook) answers part of question 2, and ADR 0035's Consequences answer part of question 1 (the hook budget is max-shaped, so lower process fan-out moves the budgeted wall only if the guardrails dispatcher is replaced; mods cannot take per-repository configuration).
Agent Brief
Type: Task
Summary: Decide which plugin features, if any, should ship as mods, and how the marketplace versions and tests mods.
Current behavior:
Mods are deferred. Plugin behavior ships as skills and settings hooks only.
Desired behavior:
A recorded decision (an ADR superseding or amending ADR 0035) that answers the issue's three questions, each with evidence: which plugins' hook behavior a mod would do better and why; the versioning and test cadence for an early-access API; and whether the skills-versus-hooks split in the plugin philosophy changes.
Acceptance criteria:
Out of scope: