autofork

Forks for Claude Code, opencode and Codex CLI. When your session goes idle — or its context crosses a threshold — autofork runs forks: background copies of the full conversation that do work with tools, on the model you pick for them, and report back — without interrupting you, and (by default) without leaving a single visible trace in your session.

Think of a fork as a background thought your agent has while you're away: update the project journal, distill notes, groom a TODO list, re-check an assumption — running quietly in the background, with the report slipped into your agent's context on your next exchange.

forks are not skills. A skill is something the model chooses to load and follow. A fork is something the harness schedules at lifecycle moments the model never sees. A fork fires because its run_on moment happened, full stop — there is no retrieval/RAG involved. (A fork can be attached to a skill — a FORK.md next to a SKILL.md — but the skill stays a skill and the fork stays a fork: the attachment only names the fork after the skill and makes sure the fork sees the skill's instructions when it runs.)

How a fork fires

1. When a turn ends, an asyncRewake Stop hook long-polls the autofork daemon in the background without blocking your session. A turn that ends while background work is still running (a run_in_background command, a background subagent) is a stop but not an idle moment: the clock starts when that work clears — see What counts as idle.
2. When forks come due (an idle deadline elapses, or a context threshold was crossed), the daemon answers the poll with the due forks.
3. By default (v0.18, fork_runner = "headless") the hook consumes the wake itself: each fork runs as a claude -p --resume <conversation> --fork-session subprocess — a true fork of your conversation, executed outside your session, on whatever model the fork (or [fork_models]) names. Its report is spooled and delivered silently as additionalContext on your next prompt. Nothing about forks ever appears in your conversation — the same quiet semantics as opencode. (The one exception is a chain: true report that asks to continue: that one wakes the session, because a goal loop only advances when the parent sees it.)
4. Opt-in (fork_runner = "subagent"): the cache-preserving alternative. The hook exits 2, which wakes the idle session with a spawn payload; the session's own model calls the Agent tool with subagent_type: "fork" per due fork — background subagents that inherit the whole conversation with ~99% prompt-cache reuse (measured: cache_read 31,681 / cache_creation 326). The trade: the wake turn, the spawn calls and the completion relays are all visible turns in your conversation, and the forks run on your session's model (overrides don't apply).

Headless forks of an interactive session read the inherited history cache-cold (Claude Code stamps request prefixes per mode) — which cheap fork models make irrelevant; that is the default posture: quiet first, cache tricks by explicit opt-in. A subagent-mode fork inherits the session's permissions and model; a headless fork takes mode:/[fork_modes] (default acceptEdits).

Requirements

autofork v0.5+ needs a Claude Code version whose Agent tool supports subagent_type: "fork" in interactive sessions:

• Claude Code >= 2.1.161 — the fork subagent is enabled by default (recommended).
• 2.1.117 – 2.1.160 — it exists but is gated; export CLAUDE_CODE_FORK_SUBAGENT=1.
• < 2.1.117 — no fork subagent; autofork v0.5 can't run forks.

autofork doctor checks your claude --version against these thresholds.

If wakes report the fork type unavailable

Even on a fully current version a wake can report Agent type 'fork' not found — the fork subagent ships behind a staged server-side rollout. The confirmed fix is to force-enable it persistently in ~/.claude/settings.json:

{ "env": { "CLAUDE_CODE_FORK_SUBAGENT": "1" } }

(Prefer this over a shell export so every session gets it.) As a safety net, each wake also tells the model to retry the fork call once and, if it still fails, to hold the spawn instructions and run them on your next message rather than substituting a wrong agent — so a transient miss self-corrects even without the pin. (Deferred agent rosters that key off the user's prompt are plausible and were briefly suspected here, but the evidence was confounded — see below — so the env pin, not any disclosure mechanism, is the remedy.)

Never let a wake create a fork agent file. If the fork type is missing, the correct fix is the env pin above — not a custom ~/.claude/agents/fork.md. A custom agent named fork does not inherit the conversation (only the built-in type does) and shadows the real one, so its "report" will show no knowledge of your session. Wakes are instructed never to create one; if you suspect an impostor slipped in (a fork "ran" but its report is context-blind), run autofork doctor — it flags fork.md under .claude/agents/. Delete it.

Install

From the plugin marketplace:

/plugin marketplace add TheUnderdev/autofork
/plugin install autofork@autofork

On first use a bootstrap step downloads the prebuilt binary for your platform from GitHub Releases into the plugin's persistent data directory (or builds it with cargo if no artifact matches). macOS (arm64/x64) and Linux (x64/arm64) are covered.

For local development: claude --plugin-dir ./plugin inside this repo.

The binaries are also on crates.io — cargo install autofork autofork-daemon — but note that installs only the CLI and daemon, not the plugin (hooks, marketplace wiring); the plugin install above is the supported path.

Writing forks

Forks are discovered upward from your project directory and at the user level, from three kinds of places:

• .autofork/forks/ trees (autofork's own layout), plus the user-level ~/.autofork/forks/
• .claude/forks/ trees — a forks/ dir next to your skills dir — plus ~/.claude/forks/
• skill folders: a FORK.md next to a SKILL.md inside .claude/skills/** or .agents/skills/** (and the user-level ~/.claude/skills/, ~/.agents/skills/ — codex's native skills location) defines a fork named after the skill — see Skill-attached forks below. Roots are deduped by canonical path, so the common setup where ~/.agents/skills and ~/.claude/skills are symlinks to one shared tree discovers each fork exactly once — no double firing.

Project definitions win name collisions over user-level ones (nearest first). Inside a forks root, two layouts mix freely (subfolders are just organization):

.autofork/forks/
├── journal.md # a fork named "journal"
├── style-guide.md # a companion NOTE (no `fork: true`) — not a fork
├── maintenance/
│ └── groom-todos.md # a fork named "groom-todos"
└── deep-review/
└── FORK.md # a fork named "deep-review"

A fork is a markdown file whose frontmatter carries fork: true: YAML frontmatter for when, body for what to do.

---fork: truedescription: Keep NOTES.md current with what happened this sessionrun_on:
- idle: 15mthrottle: 30m---
Review the session so far and update NOTES.md with any durable decisions,
open questions, and next steps. Keep it under 200 lines.

Since v0.5, .autofork/forks/ may hold arbitrary companion .md files (reference material a fork's body tells it to read, for instance). Only files marked fork: true are forks; anything else is skipped. As a guard rail, a file that looks like a fork (carries run_on, throttle, tags, after, overlap, description, …) but lacks the marker produces a warning in autofork forks, so a missing marker can't silently disable a real fork. fork: false is an explicit, silent opt-out.

Frontmatter reference

Key	Values	Default
fork	true — required on every fork	—
description	free text, for humans (autofork forks)	—
run_on	list of moments, see below	[idle]
throttle	min gap between runs: 30m, 2h, 90 (seconds)	none
after	fork name(s) to run after: journal, [a, b]	—
priority	ordering weight (z-index): lower spawns earlier, higher waits for the lower waves; equal = together	0
overlap	true to allow two runs of this fork at once	false
tags	labels for the enable/disable filter: ci, [ci, review]	—
chain	true — a run may request another by ending its report with <<autofork:continue>>	false
chain_limit	max chain runs within one pause	config chain_limit (25)
gate	true — hold the other idle forks while this fork's run/chain is unsettled	false
model	model for this fork's runs: a value or a fallback list ([sonnet, haiku] — a failed run retries on the next), scalar or keyed by client (claude-code: / opencode: / codex:)	config [fork_models], else inherit the session's
mode	operation mode for the runs (permission mode / codex sandbox / opencode agent), scalar or client map	config [fork_modes], else the client default

model: and mode: (v0.17) exist because a fork rarely needs the parent session's expensive model: your session runs on the big model, the journal fork runs on a cheap one. Fork files are shared across clients and a model id rarely means anything to more than one of them, so the map form names each client explicitly:

model:
claude-code: [sonnet, haiku] # fallback list: a failed run retries on the nextopencode: github-copilot/gemini-3.7-flashcodex: gpt-5.6-lunamode:
codex: workspace-write

A scalar applies wherever the fork runs; a client the map doesn't name inherits the session's model. Caveat: the Claude Code subagent runner cannot apply either key (fork subagents always inherit the parent session's model and permissions) — they take effect on opencode, codex, and Claude Code's headless runner below.

Moments for run_on:

• idle — the session has been quiet for the default idle deadline (config, 10m)
• idle: 20m — a custom idle deadline; idle: 0s fires at the pause's very first Stop (the goal-fork recipe below)
• context_tokens: 150000 / context_used: 80% / context_left: 20000 — context-size thresholds, each firing at most once per session
• every: 1h — at least this long since the fork's last run (before the first run: since the session began), without waiting for a pause: it fires at the first turn boundary past the interval, however brief the pause — and in opencode sessions it fires even mid-run (the plugin keeps a poll parked while the session is busy), so hour-long autonomous runs still get their periodic forks. It is a backstop for activity, not a cron: once the session goes quiet it fires at most once more (only if the last run predates the pause), then stays silent until your next genuine activity re-arms it — a session left idle overnight runs nothing. Combine with idle: for "on a 4-minute pause, or hourly regardless": run_on: [idle: 4m, every: 1h] — an idle-triggered run resets the hourly clock (and usually absorbs that one post-pause fire too).
• changed: <glob> (v0.24) — a watched path was written, created or deleted. The first external moment: it fires from the outside world, not from the session's own lifecycle, so a fork can react to another session's work, a git pull, or a build artifact landing. See External moments.
• event: <name> (v0.24) — someone ran autofork emit <name>. The same external moment for things that are not files.

Unknown keys are ignored; invalid values warn and fall back to defaults (autofork forks shows the warnings). Fork bodies should be idempotent — a fork may fire on any idle pause.

Once per pause. An idle-triggered fork fires at most once per idle pause (restoring the pre-v0.5 "fires once per idle pause" semantics). A pause is the quiet stretch after one of your turns; genuine activity starts a new one. This matters because each wake turn — and each fork-completion relay turn — ends with its own Stop, which re-arms the machinery; without the per-pause rule a fork whose throttle is shorter than its idle deadline would wake you again every cycle, forever. So within a single pause a fork issues one wake and no more, regardless of throttle; throttle still applies across pauses. (context_* thresholds are separately once-per-session.)

What counts as genuine activity: your own prompts, and any background task finishing that autofork didn't spawn (a run_in_background command, a workflow, an agent of your own). The daemon records every fork spawn's tool-use id from the session transcript, so when a completion notification arrives it can tell its own forks (a continuation of the same pause — never re-fires anything) from other background work (the session picked real work back up, so the next quiet stretch is a new pause and idle forks fire again).

after sequencing is daemon-enforced: a wake spawns only the root fork(s) and names the held dependents; the daemon keeps the dependents until it observes every predecessor's completion notification, then answers the very next Stop with their spawn instructions (telling the model to carry the predecessors' reports into the dependents' prompts — after: [research, lint] waits for both). Held dependents are dropped when you send a real message before the chain finishes (the whole chain simply re-fires on the next pause) and when the session ends. Dependencies resolve within one due batch: after sequences forks that come due together, it does not delay a fork until some other fork eventually runs.

priority orders forks that come due together without naming them: the batch runs in ascending priority waves — a wave spawns once every fork of the lower waves has finished — and forks sharing a priority spawn together. Use priority: 100 for "run this fork last no matter what else is defined", -10 for "before everything". It is enforced the same way as after (higher waves are held by the daemon and released on the lower forks' completions), but the gate is order-only — no reports are piped. after wins over priority: a dependent's effective priority is lifted to at least its predecessors', so it can never jump ahead of something it must run after.

By default two runs of the same fork never overlap: the wake block for a fork tells the model to skip spawning it if a previous run of that fork is still among its running background tasks. Set overlap: true to drop that line and allow concurrent runs.

Skill-attached forks

A skill folder holding both a SKILL.md and a FORK.md defines a fork named after the skill. Use it for background duties that are really "apply this skill when the moment comes": the fork body can stay a one-liner because the spawn prompt tells the fork to load the skill first if it isn't already in its inherited context, then follow the fork body.

.claude/skills/changelog/
├── SKILL.md # the skill, as usual
└── FORK.md # fork: true + run_on — "apply the changelog skill to this session"

The same frontmatter keys apply; autofork forks shows the linked skill.

Chain forks: the fork decides whether to run again

A fork with chain: true is told, in its spawn prompt, about the continue sentinel: if its report carries, on a line of its own, the marker

<<autofork:continue>>

autofork runs the fork again once the parent session has digested the report. Detection is maximally liberal (since v0.19.1): the marker counts anywhere in the report — its own line, mid-sentence, decorated, wherever — because models routinely ignore "a line of its own" and a missed sentinel silently ends a goal loop. The decision is made per run, by the fork itself — a run whose report contains no occurrence of the marker ends the chain, so a settling fork must omit it entirely. That turns a fork into an evaluator loop: check the state of some goal against the parent's current conversation, report what's missing, and come back after the parent has seen the report; each iteration forks the parent's current context, prior reports included.

Mechanics per client:

• Claude Code — a continuing report is delivered by waking the session with it: the parked Stop hook exits 2 carrying the report, so the parent reacts to it in the turn that wake starts (in subagent mode the fork's own completion notification does the same natively). The sentinel additionally re-arms the fork's once-per-pause latch, so it fires again at the Stop that ends that turn — which is the loop. Nothing else changes: no epoch bump, so every other idle fork stays exactly as it was.
• opencode — a sentinel-carrying report is injected as a real turn (instead of the usual zero-turn no-reply message), so the parent model reacts to it; the completion frame carries continue: true and the daemon re-arms the fork the same way.

Belts: the sentinel is honored only for chain: true forks (a fork never opted in changes nothing however it phrases its report, and the daemon re-checks the definition). A chain is capped at chain_limit runs per pause (frontmatter, falling back to the chain_limit config key, default 25). Your own next message always ends the chain — genuine activity starts a new pause and the fork re-evaluates on the next one.

Runaway protection

The per-pause counters above assume the daemon can tell genuine user activity from autofork's own turns. A duplicated client event stream can defeat that assumption — observed live with an opencode bug that left several agentic loops running for one session after a network interruption: each zombie loop reported the chain's own injected turns as fresh user activity, every report minted a new pause (re-arming every idle fork and resetting the chain counter), and the goal fork pumped the session forever, surviving even close + resume because both opencode's session id and the daemon's state persist. Three daemon-side guards hold the line (daemon-side because the daemon is the one singleton with persistent state — client-side guards die with their plugin instance and multiply with duplicated ones):

• The runaway breaker (runaway_limit, default 30; 0 disables): a hard wall-clock cap on wakes of one fork per session per rolling hour, counted against the persisted run log — immune to pause resets, session resumes, and duplicated event streams. Enforced at selection and at chain re-arm; tripping it is a warn in the daemon log, never silent. every: triggers are exempt (their interval is an explicit contract — an every: 1m fork is allowed to be a cron).
• The chain grace window (AUTOFORK_CHAIN_GRACE_SECS, default 20s): a waking prompt on an opencode session arriving right after a chain continue is treated as the chain's own injected turn, not user activity — deduping the duplicated observers' reports of that turn.
• The daemon-side overlap gate: an overlap: false fork with a run still in flight (a spawn the registry hasn't seen go terminal) is skipped at selection, whatever client-side gates think. A spawn whose terminal status was lost to a crash stops blocking after AUTOFORK_OVERLAP_SPAWN_MAX_AGE_SECS (default 30m).

Goal forks: gate: true

A goal fork combines the pieces: fire immediately after every one of your turns, keep working while the goal isn't met, and keep the consolidation forks out of the way until it's done.

---fork: truedescription: Drive the session's stated goal to completionrun_on: [idle: 0s]chain: truegate: true---
Look at the parent conversation's current goal. If it is not yet met: do the
next concrete chunk of work (or tell the parent exactly what to do next in
your report) and end your report with the continue line. If the goal is met,
or there is no active goal, report one line and stop — no continue line.

It loops autonomously on every client: the fork's continuing report is delivered as a turn the parent reacts to right away (Claude Code wakes the session with it from the Stop hook, codex blocks-and-injects, opencode injects a real turn), the parent does the work, and the fork re-evaluates at the stop that follows. It starts counting only when the session is genuinely idle, so a goal fork never evaluates while a background command it asked for is still running (see What counts as idle).

gate: true holds every other idle-triggered fork while this fork's run/chain is unsettled — they are dropped at selection without consuming their once-per-pause latches, and after-held dependents stay held. When the chain settles (a run without the sentinel, a failure, or the chain limit), the pause baseline resets: the held forks' idle deadlines measure from that moment, so a handover on idle: 4m fires 4 minutes after the goal work ends and captures all of it. every: and context_* triggers are deliberately not gated (a periodic backstop and a filling context window still matter mid-goal). A gate whose wake was fumbled (no spawn ever observed) lifts after a grace window (AUTOFORK_GATE_GRACE_SECS, default 180s) rather than silencing the session's forks for the whole pause; your own next message drops the gate immediately.

At close the hold does not apply. There is no settling and no poll left to release held forks through, so holding would drop them rather than defer them — the flush batch takes the whole set and puts the gate fork first instead. The end-runner is sequential, so a gate still leads and everything else still runs after it, in one batch.

Lifecycle hooks

Forks answer "run a model over this session's context at the right moment". Lifecycle hooks answer a different question: "run a command at a session's lifecycle moments" — no model, no fork, no context, no tokens. They exist for resource integrations that need to follow a session's life directly: workspace leases, seat locks, scratch-space allocation, external presence signals. Acquire on start, renew on activity, park on idle, release on end.

A hook is a markdown file under .autofork/hooks/ (per ancestor directory, nearest first, then user-level ~/.autofork/hooks/; bare <name>.md or <name>/HOOK.md, same as forks). The body is documentation only.

---hook: truedescription: keep the workspace lease aliveon: [session_start, activity, "idle: 5m", session_end]command: lease-tool touch --session "$AUTOFORK_SESSION_ID"timeout: 30s---
Renews this project's workspace lease while a session is alive, and releases
it when the session ends. The lease TTL covers crashes.

The daemon runs command through sh -c in the session's launch directory, with the context in environment variables — so renewing or releasing a lease never involves spawning a model.

Events (on):

event	fires	extra env
session_start	a session registers (startup, resume, clear — any event that opens a session)	AUTOFORK_SOURCE (startup/resume/clear/compact, when known)
resume	only a resumed session (source: resume; resumes arrive as a new session id)	AUTOFORK_SOURCE
activity	each genuine user prompt (the same signal that starts a new pause)	—
idle / idle: <dur>	the session has been idle that long — once per pause, while the session stays open and parked (bare idle uses default_idle_deadline)	AUTOFORK_IDLE_SECS
session_end	the session ended, from any path	AUTOFORK_END_REASON
changed: <glob>	a watched path was written, created or deleted — external, so it can fire mid-turn (see External moments)	AUTOFORK_WATCH, AUTOFORK_CHANGED_PATHS
event: <name>	autofork emit <name> was run	AUTOFORK_EMIT_NAME, AUTOFORK_EMIT_PAYLOAD

Every firing also carries AUTOFORK_HOOK_NAME, AUTOFORK_EVENT, AUTOFORK_SESSION_ID (the parent session id), AUTOFORK_PROJECT_ROOT, AUTOFORK_CWD, and AUTOFORK_CLIENT (claude-code or opencode).

AUTOFORK_END_REASON values: what the client reported for a clean end (Claude Code: clear/logout/prompt_input_exit/other; opencode: disposed when the instance shuts down normally, deleted when the session is deleted), or the daemon's own liveness fallbacks: gone (the client's OS process no longer exists — see below), lost (the session's parked poll dropped and the grace window expired), pruned (autofork prune), timeout (the session-timeout reaper). autofork status lists the most recent closes with their reason.

Design your integration around one honest caveat: no callback can cover SIGKILL, a kernel panic, or power loss — and if the machine dies, the daemon dies too. session_end is best-effort cleanup that makes the common paths prompt; a lease TTL plus renewal (activity, or an idle: ping) must remain the fallback that reclaims resources after a crash. That split is intentional: autofork owns the heartbeat, your lease store owns expiry.

Notes: a genuine user prompt starts a new pause, so idle: hooks re-arm exactly like idle forks; gate forks never hold hooks back (they are infrastructure, not context work); hook stdout/stderr go to the daemon log (autofork logs); a failing or timing-out hook is logged and otherwise inert. Fork-run sessions (opencode) never fire lifecycle hooks. autofork hooks lists what's discovered, with warnings.

A hook may also carry throttle: (v0.24) — a minimum gap between its runs, in the same spelling forks use. It matters most for the external moments below, which the outside world can raise far faster than a session's lifecycle ever does.

Feeds: a command that speaks into the session

Forks answer "run a model over this session's context". Lifecycle hooks answer "run a command at a session's lifecycle moments". A feed is the missing quadrant: a command that produces context.

	produced by a model	produced by a command
effect: outside world	—	lifecycle hooks (leases, locks)
effect: session context	forks (reports)	feeds

A feed is a lifecycle hook with deliver: set — same file, same discovery, same on: moments:

---hook: truedescription: keep the session current on recent handoverson:
- session_start
- "changed: ~/notes/handovers/**/*.md"deliver: contextthrottle: 30smax_bytes: 6000command: handover-brief---
Prints an index of the most recent handovers at session start, and just the
changed one when another session writes it.

deliver: takes three values:

value	what happens	cost
none (default)	stdout goes to the daemon log — the classic lifecycle hook	—
context	the block is delivered silently on the session's next prompt	no turn, invisible in the transcript
wake	the block is delivered by waking the session, which reacts to it in a turn	one turn

The output is framed exactly like a fork report (source: autofork, then feed: <name> (<event>)) so the model can tell a command's output from a model's report, and so every delivery lane's "is this autofork's own injection?" check keeps working.

Three rules make a feed writable as just print the current view:

• Empty output means nothing to say. A feed with no news prints nothing and costs the session not one token. This is the normal quiet path, not a failure.
• Unchanged output is never delivered twice (for deliver: context). The daemon hashes each delivered block per (session, feed); an identical block is skipped. So the simple implementation — print the whole current index every time — is also the correct one. deliver: wake is exempt on purpose: you asked to interrupt the session, and "the deploy failed" arriving twice is news the second time too. What bounds a wake feed is its trigger and its throttle:.
• max_bytes: truncates rather than drops (default 8000). One chatty feed must not crowd out the fork reports sharing the same lane.

Environment: a feed gets the same AUTOFORK_* variables every hook gets, plus AUTOFORK_DELIVER. External moments add AUTOFORK_TRIGGER and, per kind, AUTOFORK_WATCH + AUTOFORK_CHANGED_PATHS (newline-separated) or AUTOFORK_EMIT_NAME + AUTOFORK_EMIT_PAYLOAD.

Per-client delivery, because each harness has a different silent lane:

client	deliver: context	deliver: wake
Claude Code	spooled, delivered as additionalContext at your next prompt	the parked Stop poll exits 2 with the block
codex	spooled, delivered as additionalContext at your next prompt	the Stop hook blocks-and-injects it at the next turn end
opencode	injected as a no-reply message (it has no additionalContext lane) — zero turns, and it can land mid-run	injected as a real turn, with your model/agent pinned

One honest limit: on Claude Code and codex there is no channel into a running turn. A quiet feed therefore reaches the model at its next prompt, and a wake feed at the next turn boundary. opencode is the one client that can take an injection mid-run.

External moments: changed: and event:

Every moment before v0.24 was derived from the session's own lifecycle. These two are not: they fire when the outside world moves, and they work identically for forks (run_on:) and hooks (on:).

run_on:
- "changed: ~/notes/handovers/**/*.md"# a watched path was written/created/deleted
- "event: deploy"# someone ran `autofork emit deploy`

Both spellings work — the map form (- changed: src/**/*.rs) and the quoted-string form. Prefer the quoted form for a pattern containing a colon, which a YAML plain scalar cannot hold.

Patterns support *, ? and ** (a whole-segment wildcard). A leading ~/ expands to your home directory; a relative pattern resolves against the session's project root, so a definition can say docs/**/*.md and mean its own project. There are no character classes or brace expansion: write two patterns instead.

autofork watches by polling. Every watch_interval (default 2s) the daemon stats the files each pattern matches and diffs that against the previous sweep. It does not subscribe to filesystem events, deliberately: a poll needs no extra dependency, hits no per-platform watch limit, opens no descriptor per directory, and costs exactly what the interval says. What you give up is instant delivery — a change is noticed within one interval, and then lands at the session's next turn boundary or next prompt (see the delivery table above).

Three behaviours worth knowing:

• The first sweep of a pattern is a baseline, never a trigger. A session that starts and immediately matches five hundred files is not told all five hundred just changed.
• A burst is one trigger. Changes settle for watch_debounce (default 2s) before firing, so an editor's write-then-rename and a git pull's hundred files arrive as a single fire carrying the whole path set. A directory written to continuously still fires, at most one interval late.
• Deletions count. A file that disappears is a change — a feed that lists a directory needs to know.

Scan guards: .git, node_modules, target, .venv and __pycache__ are never walked into (a git pull still rewrites the worktree, which is what patterns point at); the walk stops 24 levels below a pattern's deepest wildcard-free directory; and a pattern matching more than watch_max_files (default 20000) is watched up to the cap and warned about in the daemon log, rather than silently degrading.

A fork woken by an external trigger is told what the trigger carried — the changed paths, or the emit payload — in its spawn prompt (capped at 40 lines), so it can act on what moved instead of re-deriving it.

Throttles and the runaway breaker apply to changed: and event: exactly as they do to idle: (unlike every:, which is exempt — an interval is an explicit contract, a filesystem is not). The queued trigger is consumed when the wake is issued, so a fork fires once per change, not once per poll until the next one.

autofork emit

autofork emit deploy --payload "build 412 is live"
autofork emit handover-written --payload - # read the payload from stdin
autofork emit ping --project # only sessions under this project root
autofork emit ping --session <id> # only that session

The non-filesystem sibling of changed:, for when the thing that happened is not a file: a job finished, a deploy landed, another tool has news. It reaches every open session on this machine whose forks or hooks listen for the name (scoped by the flags above), and prints how many it reached. Use it instead of a filesystem proxy when the producer can just say so — it is precise, instant, and needs no sweep.

CLI

autofork status # daemon, sessions, recent wakes
autofork forks # forks visible from here, with warnings
autofork hooks # lifecycle hooks visible from here, with warnings
autofork run <name> # print the spawn instruction to paste into an interactive session
autofork run --tag <tag> # print instructions for every fork carrying <tag>
autofork emit <name> # raise a named external event for every listening session
autofork logs [-f] # daemon log
autofork prune # close [stale?] sessions now instead of waiting for the session timeout
autofork doctor # install checks
autofork stop-daemon # retire the daemon (it restarts on the next event)
autofork opencode install # install the opencode bridge plugin (see "opencode support")
autofork opencode uninstall # remove it
autofork codex install # install + trust the Codex CLI hooks (see "Codex CLI support")
autofork codex uninstall # remove them

autofork run can no longer spawn a fork itself (forks are subagents of an interactive session); it prints the wake-style spawn instruction for you to paste into a live session.

Configuration

~/.autofork/config.toml, overridable per project in <project>/.autofork/config.toml:

default_idle_deadline = "10m"# bare `idle` deadline; 0 disables idle forkssession_timeout = "12h"# close sessions idle longer than thisquiet_period = "20m"# daemon self-exit after this much nothing (global only)wake_debounce = "5s"# batch near-simultaneous forks into one wake; 0 answers immediatelychain_limit = 25# default cap on chain runs per pause (see chain forks)runaway_limit = 30# hard cap on wakes of one fork per session per rolling hour; 0 disablesenable_tags = ["ci"] # default tag whitelist (see below)disable_tags = ["noisy"] # default tag blocklist (see below)
[tag_throttles] # min gap between wakes of any fork carrying a tagci = "1h"fork_runner = "headless"# Claude Code execution mode; "subagent" opts into cache-preserving visible forksflush_on_close = true# run the pause's unrun idle forks when a session ends (see below)background_hold = true# a session waiting on background work isn't idle yet (see below)background_hold_timeout = "30m"# after this, one unfinished task stops holding; 0 = hold foreverwatch_interval = "2s"# how often `changed:` patterns are swept; 0 disables themwatch_debounce = "2s"# a burst of writes settles this long, then fires oncewatch_max_files = 20000# cap on files one watched pattern may track (warned when hit)
[fork_models] # default fork model per client; a fork's own `model:` wins"claude-code" = ["sonnet", "haiku"] # one id, or a fallback list tried in orderopencode = "github-copilot/gemini-3.7-flash"codex = "gpt-5.6-luna"
[fork_modes] # default operation mode per client; a fork's `mode:` wins"claude-code" = "acceptEdits"# headless-runner permission modecodex = "workspace-write"# codex sandbox

wake_debounce gives near-simultaneous forks (idle deadlines close together, say) a moment to coalesce into a single wake with multiple spawn blocks. A prompt arriving during the window cancels the wake cleanly and stamps no throttles.

The headless runner (Claude Code)

fork_runner = "headless" — the default since v0.18 — is the opencode-style quiet mode: the parked Stop hook consumes wakes itself and runs each fork as a claude -p --resume <conversation> --fork-session subprocess — a true fork of your conversation, run outside your session. Reports are spooled and delivered silently on your next prompt (hook additionalContext: your model sees them, your transcript doesn't show them). Nothing about forks ever appears in your conversation.

A headless fork of an interactive session cannot reuse its prompt cache (Claude Code stamps request prefixes per mode), so each run reads the inherited history cold — which is why headless pairs naturally with cheap fork models ([fork_models], above): pay small-model input prices for the copy instead of burning your session's model on a journal update. Runs default to --permission-mode acceptEdits (headless runs can't answer permission prompts); set a fork's mode: or [fork_modes] for more or less. Runs also carry --settings '{"disableAllHooks":true}': a --resume restores the session-scoped Stop hook that Claude Code's own /goal installs, and inside a headless fork that hook refuses the stop — the run never ends, its report (sentinel included) is never captured, and the fork drifts into doing the parent's work. Disabling hooks in the fork also keeps a throwaway reviewer from firing your lifecycle hooks; set AUTOFORK_FORK_HOOKS=1 if a fork of yours depends on one.

One report never waits for your next prompt: a chain: true run that asks to continue. There the parent is the worker and the loop only advances once it has seen the report, so the parked hook delivers that block by waking the session with it (stderr + exit 2) instead of re-parking — the goal fast path, the async twin of codex's synchronous block-and-inject. A goal loop therefore runs autonomously under the headless runner too: fork evaluates → parent wakes and works → parent stops → fork evaluates again, until a report omits the sentinel. Settled and failed runs still spool silently. (Before v0.22 a continuing report only surfaced at your next manual prompt, so the loop advanced between your messages rather than on its own.)

fork_runner = "subagent" opts back into the pre-v0.18 behavior: the session's own model spawns fork subagents (near-total prompt-cache reuse, forks on the session's model) at the price of visible wake/spawn/relay turns. Pick it when cache reuse on the big model matters more to you than a quiet conversation.

What counts as idle

A turn can end while the session is still waiting: a run_in_background Bash command polling a deploy, a background subagent researching something. The harness calls that a stop, but nothing about it is idle — and firing a goal fork there evaluates a goal against work that hasn't landed.

So background_hold (on by default) holds the idle clock while the session has unfinished background work: those stops arm no idle deadlines and start no pause, and the clock starts at the first stop after the last such task reports completion. An idle: 4m handover then measures its 4 minutes from that moment, and an idle: 0s goal fork fires exactly then. every: and context_* triggers are not held (a periodic backstop and a filling context window still matter while you wait).

autofork's own fork spawns never count — a session isn't busy because autofork is forking it; fork ordering is what after, overlap and gate are for. And because a completion can go unseen (a server left running, a notification lost to a resume), one task stops holding after background_hold_timeout (default 30m; 0 holds for as long as the work runs). Set background_hold = false for the pre-v0.22 behavior, where every stop is idle.

Flush on close

flush_on_close — ON by default since v0.21 — closes the classic gap "I quit before the idle deadline, so the consolidation forks never ran": when a session ends, every idle fork that hadn't yet fired this pause runs immediately, executed by a detached end-runner that outlives the session — a claude -p --fork-session, codex exec fork, or opencode run --fork of the on-disk conversation, in after/priority order with report piping. Throttles, tag filters and the runaway breaker still apply, so a fork that already ran recently stays quiet. Reports go where they can: Claude Code spools them under the conversation (delivered if you resume it), codex spools them the same way, opencode runs are work-only. flush_on_close = false opts out if close-time runs aren't wanted.

Close-time runs are headless, so there is nobody to answer a permission prompt: each client's runner sets its own approvals (--permission-mode on Claude Code, sandbox flags on codex, --auto on opencode — which auto-approves only what is not explicitly denied, so the agent a fork's mode: picks keeps its own permission config). A session flushes exactly once, whether the SessionEnd hook or the daemon's own liveness check notices the close first.

Session liveness

A session is open for exactly as long as the process behind it is. Since v0.23 every hook forwards the client's pid (Claude Code exports its own as CLAUDE_PID; codex's waiter knows it; otherwise the CLI walks its ancestry past any wrapper shell) together with a start-time token that makes the identity immune to pid reuse, and the daemon checks it every 15 seconds. A client that is gone closes its session with reason gone — and flushes, so the consolidation forks still run.

This matters because the two older signals both depend on the client behaving on the way out, and both are missable:

• the SessionEnd hook has to run and finish — a client that exits first (or exits in a way that skips hooks) never delivers it;
• a parked Stop poll has to be there to lose — a session that ends mid-turn has none, and a poll process that gets orphaned instead of killed keeps the socket open, which read as a live session.

Each of those left the session [open] until the 12h session_timeout reaper, with its flush-on-close forks silently dropped. Now the OS is the authority: a parked poll also watches its client and exits with it (a fork run already in flight still finishes), an orphaned poll can no longer re-open a dead session, a session whose client is provably running is never reaped for being idle, and [stale?] in autofork status means "its process is gone" rather than a guess from idleness. Sessions carried over from a previous daemon (a reboot, a killed daemon) are closed but not flushed — those forks would be consolidating a conversation that ended who knows when.

Tag filtering

Forks can carry tags: in their frontmatter, and a session can then narrow which forks fire. The filter has two sets, an enable (whitelist) and a disable (blocklist), applied per fork at selection time:

• If any of a fork's tags is in the disable set, the fork is skipped — disable wins over enable.
• If the enable set is present and non-empty, a fork runs only if at least one of its tags is in it — so untagged forks are excluded by a whitelist.
• With neither set configured, every fork runs.

Two sources feed the filter, per key:

• Per session — the environment variables AUTOFORK_ENABLE_TAGS and AUTOFORK_DISABLE_TAGS (comma-separated), read from the Claude Code process env by the hook. Set them per project/shell to scope a session (AUTOFORK_DISABLE_TAGS=noisy claude).
• Defaults — the enable_tags / disable_tags config keys above (project layer over home layer). A session's env value overrides the config default for that key.

Per-tag throttles

[tag_throttles] maps a tag to a minimum gap between wakes of any fork carrying that tag — one shared budget for the whole group. A wake of any fork with the tag suppresses every other fork sharing it until the window passes. It composes with a fork's own throttle (both must pass) and layers per key (project entries override home).

throttle and the tag throttles are stamped at wake-issuance (when the daemon answers the poll), not at fork completion — a held after dependent stamps when its wake was issued, not when it is eventually released. (The daemon does observe fork completions in the transcript, but a spawn it never sees — a model that skipped or paraphrased the Agent call — must not unlock the throttle forever, so issuance stays the stamp point.)

Costs, caveats

• Every fork is a real model call billed to your Claude Code account. Because a fork inherits the parent prefix, the marginal cost is dominated by cheap cache reads (~99% reuse measured) plus the fork's own work. Use throttle, tight run_on lists, and autofork status to keep it deliberate.
• "Once per session" latches (context thresholds) reset when a session is resumed — Claude Code assigns resumed sessions a new id, so each resume leg counts fresh.
• The transcript-based context gauge parses an internal Claude Code format; if it changes, the context_* triggers degrade to inactive rather than erroring. The window used for context_used / context_left is 200k by default and 1M when the session's model carries Claude Code's [1m] marker (e.g. claude-opus-4-8[1m]); a gauge that exceeds the assumed window bumps it to the 1M tier as a fallback. The per-model window config was dropped in v0.5.
• A wake requires a live parked Stop hook. If the daemon dies while a session is idle, that idle opportunity is simply missed — the next turn re-arms it. A hook never wedges or errors a session.
• A session whose client process dies is closed automatically — see Session liveness above.

v0.4 → v0.5 migration

v0.5 is a breaking release that replaces headless fork subprocesses with fork subagents spawned by the session's own model.

• Add fork: true to every existing fork file (both <name>.md and <name>/FORK.md). Files without the marker are no longer treated as forks.
• Default run_on changed from [idle, compact] to [idle].
• Dropped moments.compact, session_start, session_end, manual_stop, and boot are no longer supported — they are parsed but warned and ignored, and a fork whose only moments are unsupported never fires (with a visible warning in autofork forks). Supported moments: idle, idle:<dur>, and the three context_* thresholds.
• Ignored frontmatter keys.delivery, model, allowed_tools, and permission_mode are parsed-and-ignored with a warning: delivery is native, and a fork inherits the session's model and permissions.
• Ignored config keys.claude_bin, concurrency, isolation, permission_mode, run_timeout/fork_timeout, context_window, [models], and the report/poll budgets are accepted-and-warned, then ignored. Old config files never hard-error. The new wake_debounce key is the only addition.
• Interactive-only. The fork subagent type does not exist in headless -p sessions, so v0.5 drops headless and postmortem support entirely.
• Cache economics. The old warning that an interactive parent's forks couldn't reuse its cache no longer applies — a fork subagent inherits the live conversation and reuses ~99% of the prefix.

opencode support (v0.9)

autofork also runs forks in opencode sessions — same fork files, same daemon, same schedule semantics (throttles, tags, after dependencies, once-per-pause idle latching). Both opencode 1.x and opencode 2 are supported by the same plugin (as of v0.15.2 — earlier versions hang silently on opencode 2). Install the bridge plugin once:

autofork opencode install # writes ~/.config/opencode/plugin/autofork.js

then restart opencode (plugins load at instance start). autofork opencode uninstall removes it; autofork doctor reports whether the installed copy is current.

Note for opencode 2: sessions are commonly hosted by a long-lived background server (opencode serve / the shared daemon the CLI attaches to), and each resident process stays frozen at the plugin version it loaded at startup. After autofork opencode install, restart every opencode process — the TUIs and any background opencode serve daemons — or they keep running the old plugin.

How opencode forks run

opencode has no fork subagent, but it has something better for this job: a native session fork (POST /session/:id/fork — the engine behind opencode run -s <id> --fork) that deep-copies the whole conversation into a new session without touching the original. The plugin listens for session lifecycle events and talks to the same autofork daemon; when a fork comes due it:

1. forks your session (a full copy — the fork inherits everything you and the model have said),
2. prompts the copy with the fork instruction, pinning your session's model and agent (a forked opencode session doesn't inherit them, and cache reuse needs an identical prefix),
3. when the copy finishes, injects its report into your session as a no-reply message — no turn is spent; your model sees the report block (source: autofork) on your next exchange,
4. reports the completion to the daemon, which releases any after dependents.

Fork-run sessions are titled autofork/<fork> (<trigger>) in the session list while they run, and are deleted automatically once the report is delivered — each run is a full copy of your conversation, and left around they silt up opencode's database at one session per fork per pause. Failed runs stick around so you can read what went wrong; a sweep at instance start removes any leftovers (failures, crashes, sessions from older autofork versions) untouched for an hour. Set AUTOFORK_KEEP_FORK_SESSIONS=1 in opencode's environment to keep every run's session instead.

every: triggers get their strongest form here: the plugin parks a poll even while the session is busy, so an every: 1h fork fires in the middle of an hour-long run — the fork copies the conversation as it stands mid-run, and its report is injected as a message the in-flight run picks up on a later step. (On Claude Code, every: fires at the first turn boundary past the interval — there is no mid-turn hook.)

Cache economics on opencode

Measured with byte-level request diffing (same methodology as the Claude Code numbers below), an opencode fork of an interactive TUI session reuses ~100% of the parent's cached prefix (e.g. cache_read 29,717 / cache_creation 535 on a live run). opencode builds identical request prefixes in every mode — TUI, opencode run, server — so the mode-stamping problem that makes Claude Code interactive parents cache-cold for subprocess forks does not exist there. One caveat: opencode requests use Anthropic's plain 5-minute ephemeral cache (no 1-hour TTL), so a fork only reuses the parent's cache when it fires within ~5 minutes of the parent's last request — keep idle deadlines short (the default fits) or budget a cold prefix write for late forks.

Requires opencode >= 1.18 (the plugin uses the v1 plugin API and the session fork route).

Codex CLI support (v0.16)

autofork also runs forks in OpenAI Codex CLI sessions — same fork files, same daemon, same schedule semantics. Install once:

autofork codex install # merges hooks into ~/.codex/hooks.json and trusts them

then start a new codex session. autofork codex uninstall removes the hooks; autofork doctor reports whether they are installed, current, and trusted.

Trust matters: codex silently skips hooks it doesn't trust. autofork codex install writes the hooks and records their trust hashes (through the same RPCs the codex TUI's /hooks command uses), so a plain hooks.json edit is never enough — always go through the installer. If your codex binary or the autofork binary moves, re-run it; autofork doctor flags both conditions. Managed environments that set allow_managed_hooks_only disable user hooks entirely — autofork cannot run there.

How codex forks run

Codex has lifecycle hooks shaped like Claude Code's, but they run synchronously — a slow Stop hook blocks the session — so autofork parks nothing in them. Instead the SessionStart hook spawns a small per-session waiter process that tails the session's rollout file (codex records turn boundaries, token usage and the model's real context window there) and talks to the autofork daemon. When a fork comes due it:

1. forks your conversation with codex's native thread fork (codex exec fork) — a new thread that inherits the full history without touching your session,
2. runs the fork instruction as that thread's first turn, headless, with your session's model and a sandbox matching your session's permission mode,
3. spools the report with the daemon; the UserPromptSubmit hook delivers it silently as additionalContext on your next prompt — your model sees the report block (source: autofork), your transcript shows nothing, and no turn is spent reacting to it (the same quiet delivery as Claude Code's headless runner; before v0.19.2 reports rode codex's message queue, which drains as a synthetic user turn the model then answers),
4. reports the completion to the daemon, which releases any after dependents; the fork's thread is then deleted (failed runs are kept for inspection — AUTOFORK_KEEP_FORK_SESSIONS=1 keeps everything).

Chain forks work unchanged: a report carrying the continue sentinel re-arms the fork after your session digests it (a continuing goal iteration is still injected as a same-turn continuation — see the fast path below).

The goal fast path (codex Stop hook)

Codex Stop hooks run synchronously and may block-and-inject: a hook that answers {"decision": "block", "reason": …} has its reason recorded as a continuation prompt, and the model reacts in the same turn. autofork's codex Stop hook uses exactly that for goal forks: when a chain: true fork is due at the pause's very first Stop (the idle: 0s goal recipe), the hook runs it right there — the session deliberately holds while the fork evaluates — and injects the report as the continuation. The parent reacts immediately: a true autonomous goal loop with zero dead time, on every codex version. Anything that isn't an idle: 0s chain fork exits the hook instantly and stays on the waiter path.

Cache economics on codex

A codex thread fork gets a fresh thread id, and codex keys the OpenAI prompt cache on it — so a fork run reads the inherited history cold (measured: cached_input_tokens: 0). That is the default and it mirrors opencode's semantics: every run is a plain native fork. If you want the cache back, opt in with AUTOFORK_CODEX_CACHE_COPY=1 (in codex's environment): a run that uses the parent's model and whose rollout is a self-contained plain-JSONL file is then executed as a cache copy — the rollout is copied into a throwaway CODEX_HOME keeping its session id and resumed there. Same id → same prompt-cache key → the parent's warm prefix is reused (~93% measured), and the parent's real home is never touched. The preflight fails closed to the native fork (compressed rollouts, paginated history, reference-backed forks), and a fork on a different model is a different cache anyway, so it always uses the native fork.

Requires codex >= 0.148 (lifecycle hooks, codex exec fork, the queue RPC, and Stop-hook blocking — all verified against that release).

Other tools

The fork file format is deliberately tool-agnostic; autofork is the reference implementation for Claude Code, opencode and Codex CLI. Other agent harnesses are welcome to read the same fork definitions natively — the format spec above is the whole contract. A harness with its own lifecycle may honor extra keys or moments as extensions (autofork warns about and ignores keys like delivery that only make sense elsewhere), and the reverse holds here: a definition written for such a harness degrades gracefully under autofork.

License

MIT