sync-opencode

Quick Summary

Goal: Reconcile the project-root opencode.json with the framework's recommended opencode defaults, compile .claude/settings.json into a portable opencode plugin (.opencode/plugins/easy-claude-hooks.js), and verify both. This runner is the single and only entrypoint for the opencode surface pipeline.

PORTABILITY CONTRACT — .claude/ and .opencode/ are portable, self-running and self-testing. Copy them into ANY repository — a Python repo, a .NET repo, a repo with no package.json at all — and the sync / verify / test entrypoint still works, because it is a path INSIDE the bundle invoked with plain node. Never drive this framework through a host package.json script, and never document one: npm run … names a command that does not exist in most projects the bundle is copied into. The only external requirement is node >= 18 — no node_modules, no npm, no lockfile. The generated plugin imports only node: built-ins and spawns the canonical .claude/hooks/*.cjs scripts.

Discover the roster instead of memorizing it: node .claude/skills/sync-opencode/scripts/run-opencode-sync.mjs --list-stages

Summary:

• opencode has no shell-command hook system — hooks are JavaScript plugin callbacks. This skill compiles .claude/settings.json into a bridge plugin whose runtime drives the original Claude hooks from opencode's plugin events.
• Recommended opencode config is part of the framework. .opencode/opencode.recommended.json is the single source of truth for the framework's opencode defaults; the config stage deep-merges it into the project-root opencode.json (recommended keys win, project-only keys survive).
• Scope is hooks + recommended config + skill permissions + the sub-agent mirror. opencode already auto-discovers skills from .claude/skills and .agents/skills, so there is no skill mirroring here (unlike /sync-codex). opencode ignores disable-model-invocation, so the skills stage enforces the selection policy through permission.skill in the project-root opencode.json instead, and writes a .opencode/commands/<name>.md per hidden skill so its explicit /name keeps working — see Skill permissions. Sub-agents are NOT auto-discovered, so .claude/agents/*.md IS mirrored into .opencode/agent/*.md by the agents stage — that is what lets the workflow protocols dispatch the same specialists (architect, code-reviewer, security-auditor, …) on opencode.
• Keep .claude canonical: edit .claude/settings.json / .claude/hooks/** and re-run this pipeline; never hand-edit the generated .opencode/plugins/easy-claude-hooks.js.
• To change a default opencode setting, edit .opencode/opencode.recommended.json, then re-run this pipeline to propagate it into the root config of every project the .opencode/ folder is copied into.
• The legacy hand-written .opencode/plugins/notification.js is superseded by the generated bridge; the runner backs it up under tmp/opencode-legacy/ and removes it so notifications are not sent twice.

Workflow:

1. Config — node .claude/scripts/opencode/sync-config.mjs deep-merges .opencode/opencode.recommended.json into the project-root opencode.json.
2. Sync — node .claude/skills/sync-opencode/scripts/run-opencode-sync.mjs rewrites permission.skill in the project-root opencode.json together with its ownership ledger .opencode/skill-permissions.generated.json, writes a marked .opencode/commands/<name>.md per hidden skill, and regenerates the bridge plugin, the .opencode/agent/*.md mirror, and the sync report.
3. Test — the runner executes the opencode tooling tests (config + writer + generated-plugin runtime).
4. Verify — the runner re-merges/re-renders in memory with the REAL writers and byte-compares against the tracked files.
5. Inspect — on failure, re-run the failing stage with --only=<stage> --verbose.

Key Rules:

• MUST run stages in order — the orchestrator fails fast on the first non-zero exit
• NEVER hand-edit .opencode/plugins/easy-claude-hooks.js; regenerate it from .claude/settings.json
• NEVER hand-edit a project-root opencode.json as the way to change framework defaults — edit .opencode/opencode.recommended.json and re-run the pipeline
• .opencode/opencode.recommended.json MUST NOT be named .opencode/opencode.json; opencode auto-loads that path as project config
• The generated plugin and .opencode/plugins/** are the only opencode hook surface; no skill mirror is produced (skills are auto-discovered, sub-agents are mirrored into .opencode/agent/**)
• Only node "$CLAUDE_PROJECT_DIR"/... hook commands are compiled; other command shapes are reported as unsupported-command-shape
• Claude events opencode cannot reproduce are reported as skipped-events in the sync report — never silently dropped
• The skills stage changes only permission.skill keys recorded in .opencode/skill-permissions.generated.json; a user-set key is never adopted, overwritten or loosened, and a called skill is never hidden; it rewrites or deletes only .opencode/commands/*.md files that carry its generated marker
• Idempotent — re-running the sync produces byte-identical plugin, config, permission, and agent output

Contents

• Quick Summary
• Why a bridge plugin (not a config mirror)
• Recommended opencode config (source of truth)
• Skill permissions
• Hook mapping
• Known limitations
• Usage
• Stages
• Closing Reminders

Why a bridge plugin (not a config mirror)

The config stage above reconciles framework-recommended opencode defaults; it is not a mirror of Claude settings. This section explains why the project's hook surface is generated as a plugin.

Codex supports lifecycle hooks and can be driven by .codex/hooks.json. opencode cannot: its only extension point for lifecycle behavior is a JS/TS plugin under .opencode/plugins/ (or an npm plugin). So instead of transcribing commands into a JSON file, this pipeline emits a self-contained plugin that spawns the canonical Claude hook scripts and translates both directions:

Direction	Translation
opencode → Claude hook	event name, tool id → Claude matcher name, opencode args → tool_input field names
Claude hook → opencode	exit 0 = allow · exit 2 = block (throw) · stdout hookSpecificOutput.updatedInput = argument rewrite · stdout hookSpecificOutput.additionalContext = injected context · stdout hookSpecificOutput.permissionDecision: "deny" = block

Recommended opencode config (source of truth)

The framework ships recommended opencode defaults as a portable template and reconciles them into whatever project the .opencode/ folder is copied into.

Item	Path
Source of truth (edit this)	.opencode/opencode.recommended.json
Generated target (never hand-edit for defaults)	<project-root>/opencode.json
Writer / verifier	.claude/scripts/opencode/sync-config.mjs (--check for verify)

To change a default recommended opencode setting in the future, edit .opencode/opencode.recommended.json and re-run /sync-opencode. Every project that receives the .opencode/ folder then gets the updated default the next time the pipeline runs. The recommended file is deliberately NOT named .opencode/opencode.json because opencode auto-loads that path as project config — keeping the .recommended.json name makes it a template, not an active config.

Merge semantics: the writer deep-merges the recommended defaults into the existing root opencode.json. Recommended keys win at every leaf; object keys that exist only in the project survive untouched; arrays in the recommended file replace the project's array. A project with no root config receives the recommended defaults verbatim. A malformed existing root config is reported, never clobbered.

Compaction: host default

The framework pins no auto-compaction budget on any of its three surfaces (Claude Code, Codex, opencode), so each host applies its own default and each user sets their own. The recommended file therefore carries no model limit: opencode takes the window from the model's registry entry (models.dev). For the pinned model that is context: 1000000, output: 384000 (verified with opencode models opencode-go --verbose, v1.18.31).

opencode has no absolute compaction threshold — it compacts relative to the model's declared window, so the window IS the knob. From session/overflow.ts (v1.18.31):

usable = limit.input ? max(0, limit.input - (compaction.reserved ?? min(20_000, maxOutput)))
                     : max(0, limit.context - maxOutput)
maxOutput = min(limit.output, 32_000)          // OUTPUT_TOKEN_MAX
compaction happens once total tokens >= usable

So the registry window compacts at 968,000 tokens (1,000,000 − 32,000). A user who wants an earlier point sets their own limit on the model — in the project-root opencode.json or their global opencode config. For example, limit.context: 500000 + limit.output: 384000 compacts at 468,000.

Retiring the old pin: the config stage removes the model's limit from the root opencode.json only when it is exactly the formerly bundled { "context": 500000, "output": 384000 } (the deep merge alone never deletes a key). Any other limit is the user's: it is kept and reported with one kept user-set … line. Keep a personal 500K budget in the global opencode config if the root file must stay untouched.

Two traps for anyone setting their own limit, both verified against the released source:

• compaction.reserved is inert here. It is only read on the limit.input branch, and models.dev declares no input for this model. Raising it does nothing.
• Do NOT express the budget as limit.input. It is undocumented, and limit.context is what the rest of opencode reads as the window — the TUI context percentage, ACP usage reporting, and the limits handed to the AI SDK. Capping input while leaving context at 1M shows ~48% in the TUI at the moment it compacts.

Copying the framework into a new project: copy .claude/ and .opencode/ (including .opencode/opencode.recommended.json), then run /sync-opencode — it generates/updates that project's root opencode.json, bridge plugin, and reports. Do NOT copy .opencode/skill-permissions.generated.json or .opencode/commands/: both are generated per project by the skills stage, and a copied ledger is ignored anyway because it names the other project — why: a ledger from another project would otherwise claim the adopter's own permission.skill entries.

Skill permissions

opencode lists every discovered skill to the model and ignores disable-model-invocation, so the skills stage (.claude/scripts/opencode/sync-skills.mjs) writes the framework's selection policy into permission.skill of the project-root opencode.json. A deny entry hides the skill from the model and rejects loading it through the skill tool.

Skill	Effective tier or mark	permission.skill entry
Command-only skill (disable-model-invocation: true)	—	deny
Workflow wrapper (skill name = a .claude/workflows.json id)	manual	deny
Workflow wrapper	confirm	ask
Workflow wrapper	auto	none
Skill in skillProfile.nameOnly	—	no entry from the profile; a stricter entry from a row above stays
Skill in skillProfile.commandOnly or skillProfile.off	—	deny plus a generated command, so /name still runs it
Any other skill	—	none

• Effective tier, team scope. A wrapper follows its effective tier — the project override for that workflow, otherwise the stricter of its framework tier and the project default (portability.workflowActivation in the project config, default docs/project-config.json) — never the raw workflows.json value. The tier wins over the wrapper's own disable-model-invocation mark, so an override that loosens a manual workflow to auto removes its entry. opencode.json is a project file, so a developer's .claude/.ck.local.json never lands in it.
• Called skills are never hidden. A skill named as a step of any workflow (sequence or any variants.*.sequence), in an agent's skills: frontmatter list, or in the curated calledByOthers or entrySkills lists of .claude/config/skill-profiles.json gets no policy entry, and the run prints one line per skill: skipped <name>: called by <callers>. The called set has one owner, resolveProfile().called in .claude/scripts/sync-skill-profile.cjs. A .claude/workflows.json without a workflows map stops the stage instead of counting as empty.
• Skill profile. The skillProfile rows come from the same resolveProfile() as the Claude sync (read .claude/config/README.md → Skill profile when you need the presets and lists). nameOnly writes nothing, because an ask would stop every workflow step that loads the skill; it never loosens an entry the policy already writes. A called skill in commandOnly or off is refused unless skillProfile.allowHidingCalledSkills is true: the stage prints the resolver's message and skill-profile: nothing was written, exits non-zero, and --check fails the same way. With the opt-in the deny is written and a warning line names the skill. Resolver warnings print as warning: skill profile: ... only when skillProfile is declared. Profile entries go through the same ownership ledger, so a user key is never adopted, overwritten or loosened.
• Ownership. .opencode/skill-permissions.generated.json lists each key the generator owns, with the value it wrote. Only listed keys are updated or removed.
• The ledger is bound to its project. It records project = project.name from the project config (default docs/project-config.json; .claude/.ck.json portability.projectConfigPath relocates it), or null when there is none. A ledger whose project is missing or differs — copied from another project, or left over from a rename — is ignored with one line, warning: ignored .opencode/skill-permissions.generated.json: it belongs to another project (...), owns nothing, and is rewritten for this project, so every entry already in opencode.json stays the user's. A project config that exists but is not valid JSON stops the stage before anything is written.

Situation	Outcome
Key existed before the generator first wrote it	Kept and never recorded as owned; conflict: permission.skill.<name> is <user value>, generator wants <value>; kept the user value when it differs
Key listed only in a ledger from another project	Treated as the previous row: never owned
Owned key the user changed	Kept; the same conflict line
Owned key the user deleted	Restored to the policy value with no conflict line — set an explicit value such as allow to keep a skill loadable
permission or permission.skill is a single value, not a map	Unchanged; one conflict line; no skill entries written
Wildcard key such as internal-*	Always the user's; never rewritten

• Only the skill tool is restricted. The stage writes no read, edit or other permission, so reading .claude/skills/<name>/SKILL.md by path keeps working for workflows that load a skill file directly.
• A skill folder whose name is not lowercase letters, digits and hyphens (^[a-z0-9][a-z0-9-]*$) is skipped with a warning.
• --check fails when permission.skill, the ledger or a generated command differs from a fresh sync. A kept user value or user command is not drift.

Commands for hidden skills

A hidden skill keeps its explicit /name: the same stage writes .opencode/commands/<name>.md for every skill whose final exact permission.skill entry is deny — the policy's own entries and a user deny alike (wildcard keys are not evaluated). A skill the policy denies but whose kept user value is not deny, or whose entry could not be written because permission / permission.skill is not a map, gets no command.

---
description: "<the skill description, whitespace collapsed, YAML-escaped>"
---

<!-- GENERATED OPENCODE COMMAND (sync-skills.mjs) for .claude/skills/<name>/SKILL.md — do not hand-edit; re-run:
     node .claude/skills/sync-opencode/scripts/run-opencode-sync.mjs -->

@.claude/skills/<name>/SKILL.md

Arguments: $ARGUMENTS

• Name. The command name is the skill folder name, never the name: inside the skill; the write path must resolve inside .opencode/commands/.
• Marker ownership. Only files carrying the GENERATED OPENCODE COMMAND marker are rewritten or deleted. A marked command whose skill is no longer hidden is removed.
• Same-name user command. A command without the marker that already uses a hidden skill's name — in .opencode/commands/ or opencode's singular .opencode/command/ folder — is kept unchanged, no generated command replaces it, and the run prints conflict: <path> is a user command without the generated marker; kept it, no command generated for skill <name>.
• No shell in generated bodies. Generated commands never contain !`…`. Shell note: opencode substitutes $ARGUMENTS before it runs !`cmd` injections, so typed arguments that themselves contain !`…` run as shell on that host. Do not paste untrusted text as command arguments.

Hook mapping

Claude hook event	opencode plugin surface	Notes
PreToolUse	tool.execute.before	throw blocks the tool; updatedInput rewrites output.args
PostToolUse	tool.execute.after	additionalContext is appended to the tool output
UserPromptSubmit	chat.message	additionalContext is injected as a synthetic text part
SessionStart	event:session.created / event:session.compacted + experimental.chat.system.transform	additionalContext is injected into the system prompt
SessionEnd	event:session.deleted
Stop	event:session.idle
Notification	event:question.asked / event:permission.asked	matcher vocabulary AskUserPrompt / permission_prompt
PermissionRequest	permission.ask	blocked → status: "deny"

Tool id aliases (opencode → Claude matcher names the hooks are written against):

opencode tool	Claude matcher names
bash	Bash
edit	Edit, MultiEdit
write	Write
read	Read
grep	Grep
glob	Glob
apply_patch	Edit, Write, MultiEdit, NotebookEdit
todowrite	TodoWrite, TaskCreate, TaskUpdate, update_plan
webfetch / websearch	WebFetch / WebSearch
question	ask user question tool
skill	Skill
<server>_<tool> (MCP)	mcp__<server>__*

Known limitations

• MCP argument visibility. opencode registers MCP tools as <server>_<tool>, and Claude matchers like mcp__github__* are matched against that convention. However, opencode does not expose MCP tool arguments to tool.execute.before, so an MCP PreToolUse hook that inspects tool_input cannot see them on this host.
• apply_patch has no file_path. The bridge reports it as Edit/Write/MultiEdit and passes patchText through as patch; hooks that require file_path (doc-sync-gate) allow/ignore it exactly as they do on other hosts.
• Session resume. SessionStart runs on session.created and session.compacted; when opencode resumes a session without either event, the bridge runs source: "startup" once on the first chat.message.
• Claude-only stdout fields other than updatedInput, additionalContext, and permissionDecision are ignored by the bridge.

Usage

# Discover the stage roster:
node .claude/skills/sync-opencode/scripts/run-opencode-sync.mjs --list-stages

# Full sync + test + verify:
node .claude/skills/sync-opencode/scripts/run-opencode-sync.mjs

# Stream live child output:
node .claude/skills/sync-opencode/scripts/run-opencode-sync.mjs --verbose

# Every read-only gate (no mutation):
node .claude/skills/sync-opencode/scripts/run-opencode-sync.mjs --verify-only

# Just regenerate the plugin:
node .claude/scripts/opencode/sync-hooks.mjs

# Just verify the tracked plugin is current:
node .claude/scripts/opencode/sync-hooks.mjs --check

# Just reconcile the recommended root opencode.json:
node .claude/scripts/opencode/sync-config.mjs

# Just verify the root opencode.json is current:
node .claude/scripts/opencode/sync-config.mjs --check

# Just write the skill permissions:
node .claude/scripts/opencode/sync-skills.mjs

# Just verify the skill permissions are current:
node .claude/scripts/opencode/sync-skills.mjs --check

# Skip a stage while debugging:
node .claude/skills/sync-opencode/scripts/run-opencode-sync.mjs --skip=hooks

Exit codes: 0 all pass · 1 orchestrator failure · non-zero propagates from failing stage.

Stages

9 stages, sequential — the complete opencode surface pipeline, owned entirely by this runner:

#	Stage	Script	Effect
1	config	.claude/scripts/opencode/sync-config.mjs	Deep-merge .opencode/opencode.recommended.json into the project-root opencode.json
2	skills	.claude/scripts/opencode/sync-skills.mjs	Write permission.skill in the project-root opencode.json, the ownership ledger .opencode/skill-permissions.generated.json, and a marked .opencode/commands/<name>.md per hidden skill (see Skill permissions)
3	hooks	.claude/scripts/opencode/sync-hooks.mjs	Generate .opencode/plugins/easy-claude-hooks.js + tmp/opencode-hooks.sync.report.json; back up/remove legacy notification.js
4	agents	.claude/scripts/opencode/sync-agents.mjs	Mirror .claude/agents/*.md into .opencode/agent/*.md (mode: subagent + the canonical body verbatim)
5	tests	Runner discovers .claude/scripts/opencode/tests/*.test.{mjs,cjs}	Run opencode tooling tests; missing or empty discovery fails
6	verify-config	.claude/scripts/opencode/sync-config.mjs --check	Re-merge with the REAL writer and byte-compare with the project-root opencode.json
7	verify-skills	.claude/scripts/opencode/sync-skills.mjs --check	Re-plan with the REAL writer; fail when permission.skill, the ledger, or a generated command is missing, changed or stale
8	verify-hooks	.claude/scripts/opencode/sync-hooks.mjs --check	Re-render with the REAL writer and byte-compare with the tracked plugin
9	verify-agents	.claude/scripts/opencode/sync-agents.mjs --check	Re-render every agent with the REAL writer and byte-compare with the tracked .opencode/agent/*.md

Closing Reminders

MUST ATTENTION keep the /sync-opencode skill user-invoked-only; no unrelated skill, agent, or workflow may auto-run the mutating pipeline. MUST ATTENTION edit .claude/settings.json and .claude/hooks/** as the source, then regenerate; NEVER hand-edit .opencode/plugins/easy-claude-hooks.js MUST ATTENTION the framework's default opencode settings live in .opencode/opencode.recommended.json — edit THAT file to change defaults, then re-run this pipeline; never treat a project-root opencode.json as the source MUST ATTENTION the generated plugin must import only node: built-ins so .claude/.opencode stay portable into any project MUST ATTENTION the config stage deep-merges (recommended wins, project-only keys survive) and never clobbers a malformed root config — it reports instead MUST ATTENTION the runner auto-resolves the repo root from its own path — do not pass a cwd flag MUST ATTENTION the legacy notification.js is backed up, never destroyed, before removal

Anti-Rationalization:

Evasion	Rebuttal
"Just edit the .opencode plugin directly"	Next sync overwrites it. Edit .claude/settings.json and regenerate.
"Just edit the project-root opencode.json to change the defaults"	The next sync re-merges the recommended file and your default is lost, and no other project gets it. Edit .opencode/opencode.recommended.json and re-run.
"Make the recommended file .opencode/opencode.json"	opencode auto-loads that exact path as project config, so it would stop being a template. Keep the .recommended.json name.
"Skip the tests stage"	The tests exercise the generated plugin against real hook subprocesses; skipping ships an untested bridge.
"Mirror skills too for symmetry with sync-codex"	opencode already discovers .claude/skills; a mirror would duplicate and drift. Out of scope by design.

[FAILS FAST] First non-zero stage exit aborts the chain. Re-run the failing stage with --only=<id> --verbose. [REPO ROOT] The orchestrator auto-resolves the repo root from its own path.