@selesai/code 0.13.39 → 0.14.0

package/dist/skills/pi-subagents/references/multi-lane-orchestration.md

@@ -2,6 +2,16 @@
 
 Use this reference when several independent tasks need coordinated workers, worktrees, or repositories. It defines lane ownership; use the other pi-subagents references for run controls, prompts, and mission details. The parent remains the final decision-maker.
 
+Create lanes only after delegation is operator-authorized and each lane materially improves evidence, independent review, specialization, useful parallelism, or isolated execution. Do not manufacture parallelism: keep dependent work serial, and only split work when each lane has a distinct decision and useful output.
+
+## Mandatory implementation-topology preflight
+
+Before launching a writer for substantial mutation work, classify the implementation as **single-seam** or **multi-seam**. Two or more independently testable contracts or source boundaries make it multi-seam. One PR does not imply one writer, and the one-writer-per-worktree rule prevents concurrent ownership conflicts; it does not authorize an issue-wide commission.
+
+Record the classification and evidence on a lane board before writer launch. For multi-seam work, give each isolated component writer exclusive file or contract ownership and focused validation gates. Require every component to produce a committed or otherwise durable handoff, then start a separate integration-only owner after the component handoffs are ready. Mechanically reject and repartition a writer prompt if it owns all major outcomes across multiple independent boundaries.
+
+A single complete writer remains valid only when the board records evidence that the work is genuinely one seam or that splitting it would create overlapping ownership or artificial handoffs. This preflight must not manufacture parallelism: do not split tightly coupled work merely to increase the lane count.
+
 ## Lane board and authority
 
 Before multiple mutation-capable lanes start, record this board in the parent context:
@@ -20,20 +30,32 @@ For Pi extension repositories, keep lane worktrees outside auto-discovered exten
 
 Partition fanout by repository, source seam, decision, or review angle. Each run needs a stable key, lane-specific task, and a managed output path when a file is needed. Do not launch prompts that differ only by item name or broad file glob.
 
-Use one async `workflowScript` for a coordinated wave. Use `runs.all` for independent lanes and `runs.run` for dependent lane stages. Give cross-repository runs explicit `cwd` values and lane-qualified outputs. Use `outputMode: "file-only"` when a report must survive the run or feed a later stage. Keep scratch outputs relative so they live under subagent artifacts; use absolute paths only for durable memory, approved docs paths, or final handoff files.
+### Cold-start packets and bounded orchestration audits
+
+Every child packet must stand alone: include the goal, exact repository/cwd/ref, authority and edit boundary, relevant context/evidence, success criteria, validation, expected output, and stop/escalation rules. Do not rely on parent history, an issue number, or a broad glob alone. An orchestration audit by a top-reasoning critic model is read-only and returns at most three cited omissions; use high thinking only as an explicit parent/user escalation, never as an autonomous root or a parallel placeholder.
+
+Use one async workflow script (`subagent({ workflow: true, async: true })` after a ```` ```js workflow ```` block) for a coordinated wave. Use `runs.all` for independent lanes and `runs.run` for dependent lane stages. Give cross-repository runs explicit `cwd` values and lane-qualified outputs. Use `outputMode: "file-only"` when a report must survive the run or feed a later stage. Keep scratch outputs relative so they live under subagent artifacts; use absolute paths only for durable memory, approved docs paths, or final handoff files.
 
 ## Keep independent work moving
 
 While one lane waits, run safe independent preparation, validation, or fresh read-only review lanes. Do not block the parent just because a run is active. If no safe lane remains, record the blocker and the event that will reopen work.
 
+In an ordinary interactive session, completion wakes the parent; after useful
+async lanes are launched or triaged, yield rather than use
+`bg_wait({ all: true })` as a barrier. “Continue/orchestrate/work until
+done” means keep the board moving while safe immediate work remains. If only
+async lanes are running, record the revisit trigger and yield.
+
 An ordinary coordinated workflow has one mission. Use its durable state, artifacts, run records, and receipts for recovery. Treat a receipt as evidence, not as authority or acceptance.
 
-After a writer produces a candidate, run the required fresh-context, read-only reviewer. The reviewer inspects the exact worktree and returns evidence-backed findings. The parent decides which findings are in scope and whether the lane is ready. Send accepted fixes to that lane's sole writer, then rerun only the affected gate.
+After a writer produces a candidate, run the required fresh-context, read-only reviewer. The reviewer inspects the exact worktree and returns evidence-backed findings. The parent decides which findings are in scope and whether the lane is ready. Use `review-and-validation.md` for finding disposition, validation, and gate-failure triage. Send accepted fixes to that lane's sole writer, then rerun only the affected gate.
 
 ## Handoff, cleanup, and recovery
 
 Use stable lane-qualified artifact paths for reports and review output. A handoff states the lane status, repository and worktree, changed files, validation, open decisions, next action, and artifact or receipt paths. Copy only the final evidence to memory, a mission record, or a PR/comment, then remove scratch files from the active worktree before closing the lane.
 
+For backlog lanes and other subagent-governed workflows, a setup/runtime/tooling failure remains an infrastructure blocker. Preserve the exact failure, run/status, and repository/cwd/worktree/branch/ref state; verify a clean worktree or capture any partial diff before a same-protocol retry or asking the owner. Do not start another writer or switch to an external, foreground, or CLI execution mode without explicit owner approval.
+
 Keep a worktree until its handoff is durable, no run owns it, and no later gate needs it. Clean up only inside the recorded authority boundary. If a run stops or needs attention, preserve its worktree and artifacts, record the last known state and recovery owner, then resume that run or create one replacement lane from the handoff. Do not start another writer while worktree ownership is uncertain.
 
 Before completion, inspect the board. Every lane must be terminal or blocked with a named next action. Confirm one writer per repo/cwd or worktree, required validation, required fresh read-only review, and a durable handoff. The parent reports outcomes, evidence, residual risks, and the next decision.

package/dist/skills/pi-subagents/references/prompting-and-roles.md

@@ -8,37 +8,43 @@ Parent extensions may register a session-scoped, out-of-band ceiling through `pi
 
 ## When to Use
 
-- **Complex work orchestration**: use Fable mode as the default parent-agent loop for complex work. Complex means the task has multiple moving parts, unclear acceptance, cross-cutting code, meaningful user-visible impact, expensive or irreversible validation, broad review surface, or the user asks for orchestration. Lightweight one-off delegation can stay lightweight.
-- **Advisory review**: use fresh-context `reviewer` agents for adversarial code review, or fork to `oracle` when inherited decisions and drift matter
+All launch guidance below assumes delegation was requested by the operator in
+the current request or applicable user/project instructions. Complexity,
+workflow fit, and potential quality gains help choose a shape after that gate;
+they do not authorize a launch.
+
+- **Complex work orchestration**: after delegation is authorized, keep the parent on its ordinary strong default model and launch only when a bounded child materially improves evidence, independent review, specialization, useful parallelism, or isolated execution. For hard orchestration or root-cause questions, use a top-reasoning model only as a bounded read-only critic/oracle escalation, never as an autonomous root. Lightweight one-off delegation can stay lightweight.
+- **Advisory review**: use fresh-context `reviewer` agents for adversarial code review; fork to `oracle` only for rare escalation where inherited decisions, drift, model routing, root cause, or hard tradeoffs matter
 - **Implementation handoff**: have `oracle` advise, then `worker` implement only after an approved direction
 - **Recon and planning**: use `scout`, then write a plan when needed
 - **Parallel exploration**: run multiple non-conflicting tasks concurrently
-- **Regular skill specialists**: when discovery shows proactive skill subagent suggestions and the current work is broad enough, launch a small fresh-context fanout that asks one subagent per relevant regularly used skill to apply that skill's perspective to the task
-- **Long-running work**: launch async/background runs and inspect them later. For mutation-capable work, bound the delivery slice and elapsed runtime, then request checkpoints after active tool work returns. Reserve hard turn and tool-call caps for explicitly read-only children.
+- **Regular skill specialists**: when authorized delegation names or benefits from a relevant specialization, discovery suggestions may help select a small fresh-context fanout
+- **Long-running work**: launch async/background runs and inspect them later. For mutation-capable work, bound the delivery slice and elapsed runtime, then request checkpoints after active tool work returns. Reserve hard tool-call caps for explicitly read-only children.
 - **Subagent control**: watch needs-attention signals and soft-interrupt only when a delegated run is genuinely blocked
 - **Agent authoring**: create, update, or override project agents. Treat saved chain records as legacy inspection or migration inputs, not as a current authoring target.
 
 ## Tool vs Slash Commands
 
-Agents use the `subagent(...)` tool with `workflowScript` for execution, and `action` for management, status, and control. Humans often use the slash-command layer instead:
+Agents use the `subagent(...)` tool for execution, management, status, and control. Direct `{ agent, task }` execution is enough for one bounded child task; use a workflow script (one ```` ```js workflow ```` block in the reply plus `subagent({ workflow: true })`) when the parent needs JavaScript control flow or data-dependent branching, keyed, parallel, sequential, retry, retained-resume, aggregate, or explicit staged-lane behavior (`runs.lanes`). Humans often use the slash-command layer instead:
 
 - `/run` — launch a single agent
-- `workflowScript` — the sole public surface for sequence, parallelism, branching, retries, and aggregation
+- `subagent({ workflow: ... })` — the workflow-script surface for sequence, parallelism, branching, retries, and aggregation; with `disabledFeatures: ["workflow-scripts"]`, top-level `chain` and `tasks` replace it (see [configuration](../../../docs/configuration.md#chain-and-tasks-without-workflow-scripts))
 - `/subagents` — interactive admin for inspecting agents and editing model, thinking, or system prompt
 - `/subagents-stop [run-id]` — stop a current-session top-level async run; opens a selector when no id is given
 - `/subagents-detach [run-id]` — detach an active foreground single-subagent run without terminating its child
+- `/subagents-steer <run-id> [--child <child-id>] <message>` — steer a live async run (or one child of it) from non-TUI sessions and RPC hosts
 - `/subagent-cost` — show parent plus child token usage and cost for the session
-- `/subagents-fleet` — open the live fleet inspector with per-child controls; `Ctrl+Alt+F` opens it during an active foreground turn, `↑↓`/`jk` selects children, `PgUp`/`PgDn` scrolls transcript detail, `s` steers the selected live async child, and `D` stops its top-level async run after confirmation
+- `/subagents-fleet` — open the live fleet inspector with per-child controls; `↑↓`/`jk` selects children, `PgUp`/`PgDn` scrolls transcript detail, `s` steers the selected live async child, and `D` stops its top-level async run after confirmation
 - `/subagents-watchdog` — inspect or configure the opt-in adversarial change watchdog (model, on/off, recommend-model, check)
 - `/subagents-doctor` — diagnose setup, discovery, async paths, and intercom bridge state
 - `/subagents-models [agent]` — show the live runtime-loaded builtin model mapping
 - `/subagents-profiles`, `/subagents-load-profile`, `/subagents-refresh-provider-models`, `/subagents-generate-profiles`, `/subagents-check-profile` — manage model profiles and provider catalogs
-- `/prompt-workflow` — run a prompt template through native workflowScript execution
+- `/prompt-workflow` — run a prompt template as a native workflow script
 
 Prefer the tool when you are writing agent logic. Prefer the slash commands when
 you are guiding a human through an interactive flow.
 
-Packaged prompt shortcuts are also available for repeatable workflows. Treat them as reusable orchestration recipes, not just human slash commands. When the user asks for one of these shapes, or when the workflow clearly fits, apply the same pattern directly with `subagent(...)` and other tools:
+Packaged prompt shortcuts are also available for repeatable workflows. Treat them as reusable orchestration recipes, not just human slash commands. When the user asks for one of these shapes, apply the same pattern directly with `subagent(...)` and other tools:
 - `/parallel-review` — fresh-context reviewers with distinct review angles, then synthesis
 - `/review-loop` — parent-orchestrated worker, fresh-reviewer, and fix-worker cycles until clean or capped
 - `/parallel-research` — combine `researcher` and `scout` for external evidence plus local code context
@@ -50,19 +56,23 @@ Packaged prompt shortcuts are also available for repeatable workflows. Treat the
 
 The prompt templates in `prompts/` encode workflows the parent agent can run on demand. If the user provides a URL, issue, PR, plan, local file, screenshot, or freeform target, treat that target as the primary scope: read or fetch it before launching children, then include it explicitly in every child task. For targets outside the parent cwd, include the exact repository, explicit `cwd`, authority boundary, and expected output path in each child task. Do not depend on the parent conversation history when the recipe calls for fresh context.
 
+### Commission-risk and cold-start packets
+
+After the operator-authority gate above, delegate only when the child materially improves evidence, independent review, specialization, useful parallelism, or isolated execution; do not manufacture parallelism. Every child packet must be cold-start complete: state the goal, exact target/cwd/ref, authority and edit boundary, relevant context/evidence, success criteria, validation, output, and stop/escalation rules. For an orchestration audit by the critic tier, make the child read-only and request at most three omissions, each cited to a file, line, or decision; high thinking is an explicit escalation, not a default.
+
 ### Council Mode technique
 
-Use Council Mode when the user asks to convene advisors, debate a material decision, cross-examine recommendations, or critique and improve a plan with several model perspectives. This includes requests such as “run a council on this architecture,” “have Sol, Fable, and Kimi critique this plan,” or “get multiple oracles to debate the tradeoffs.” Read `../council-mode/SKILL.md` and follow its bounded parent-supervised protocol instead of launching ad hoc parallel oracle calls.
+Use Council Mode when the user asks to convene advisors, debate a material decision, cross-examine recommendations, or critique and improve a plan with several model perspectives. This includes requests such as “run a council on this architecture,” “have the configured advisors critique this plan,” or “get multiple oracles to debate the tradeoffs.” Read `../council-mode/SKILL.md` and follow its bounded parent-supervised protocol instead of launching ad hoc parallel oracle calls.
 
-Council advisors are read-only. User or project `council-*` profiles can pin models such as GPT 5.6 Sol, Fable, or Kimi and define any persistent stance in the profile body. Package advisors such as Surf's `gpt-pro` can join the roster only when the `surf-cli` Pi extension is installed and its `surf-oracle` provider is registered; treat them as external runners, omit child `async` for attached results, and do not pass `outputSchema` to them. The council question and scope provide the decision frame; do not invent per-advisor role labels. The parent collects independent reports, optionally sends curated cross-exam packets, and writes the final memo. Do not treat the council as agent-to-agent chat, implementation authority, or a writer swarm.
+Council advisors are read-only. User or project `council-*` profiles choose allowed models and define any persistent stance in the profile body. A top-reasoning advisor remains bounded and read-only; it does not become the root. Package advisors such as Surf's `gpt-pro` can join the roster only when the `surf-cli` Pi extension is installed and its `surf-oracle` provider is registered; treat them as external runners, omit child `async` for attached results, and do not pass `outputSchema` to them. The council question and scope provide the decision frame; do not invent per-advisor role labels. The parent collects independent reports, optionally sends curated cross-exam packets, and writes the final memo. Do not treat the council as agent-to-agent chat, implementation authority, or a writer swarm.
 
 ### Parallel review technique
 
-Use this when the user wants adversarial review of a diff, plan, issue, file, or implemented work. Launch fresh-context `reviewer` agents with distinct angles generated from the actual target. Common angles are correctness/regressions, tests/validation, and simplicity/maintainability; adapt for TypeScript, UI, security, docs, or large structural changes. Reviewers should inspect files and diffs directly, return concise evidence-backed findings with file/line references, and avoid edits unless the user explicitly asks for a writer pass. Filter on evidence, not severity: report only concrete current issues caused or made reachable by the target diff, with source proof, a test or repro, or a contract contradiction. Label findings P0/P1/P2 and end with `Merge verdict: BLOCK`, `Merge verdict: OK`, or `Merge verdict: OK with notes`. Use `blockers only` only for final pre-merge re-checks after P1/P2 findings are already captured, or for explicit emergency hotfix lanes where non-blocking findings are intentionally deferred. For targeted follow-up, ask only whether the named finding was resolved, whether the fix introduced a new defect in the fix blast radius, and whether prior P1/P2 notes still stand. For bot or PR-comment triage, classify each comment as VALID, STALE, INVALID, or OUT-OF-POLICY against current HEAD, then assign P0/P1/P2 only to VALID comments. The parent synthesizes fixes worth doing now, optional improvements, and feedback to ignore/defer before applying anything.
+Use this when the user wants adversarial review of a diff, plan, issue, file, or implemented work. Launch fresh-context `reviewer` agents with distinct angles generated from the actual target. Common angles are correctness/regressions, tests/validation, and simplicity/maintainability; adapt for TypeScript, UI, security, docs, or large structural changes. Reviewers should inspect files and diffs directly, return concise evidence-backed findings with file/line references, and avoid edits unless the user explicitly asks for a writer pass. Filter on evidence, not severity: report concrete current issues within the named review target, with source proof, a test or repro, or a contract contradiction. For a diff review, require that the issue is caused or made reachable by that diff. Label findings P0/P1/P2 and end with `Merge verdict: BLOCK`, `Merge verdict: OK`, or `Merge verdict: OK with notes`. Use `blockers only` only for final pre-merge re-checks after P1/P2 findings are already captured, or for explicit emergency hotfix lanes where non-blocking findings are intentionally deferred. For targeted follow-up, ask only whether the named finding was resolved, whether the fix introduced a new defect in the fix blast radius, and whether prior P1/P2 notes still stand. For bot or PR-comment triage, classify each comment as VALID, STALE, INVALID, or OUT-OF-POLICY against current HEAD, then assign P0/P1/P2 only to VALID comments. The parent synthesizes fixes worth doing now, optional improvements, and feedback to ignore/defer before applying anything.
 
 ### Proactive skill-specialist technique
 
-Use this when `{ action: "list" }` reports proactive skill subagent suggestions and the user's task would benefit from perspectives the parent regularly uses. These suggestions are conservative: a skill is recommended only when it is available and referenced repeatedly by configured agents or saved chains. Treat the list as an opt-in hint for the current task, not a command to always fan out.
+Use this only within operator-authorized delegation when `{ action: "list" }` reports skill subagent suggestions relevant to the requested handoff. Availability is a selection hint, not authority or a command to fan out.
 
 Default guardrails:
 - Keep the fanout small: usually one or two skill-specialist children, never more than the listed recommendations or configured cap.
@@ -73,24 +83,23 @@ Default guardrails:
 
 Example shape:
 
+```js workflow
+const results = await runs.all([
+  { key: "deslop", agent: "reviewer", task: "Apply the available 'deslop' skill to review the current diff for concrete cleanup findings only. Do not modify files.", skill: "deslop" },
+  { key: "accessibility", agent: "reviewer", task: "Apply the available 'accessibility' skill to review the UI changes for concrete issues only. Do not modify files.", skill: "accessibility" }
+]);
+return results.map(result => result.output);
+```
+
 ```typescript
-subagent({
-  workflowScript: `
-    const results = await runs.all([
-      { key: "deslop", agent: "reviewer", task: "Apply the available 'deslop' skill to review the current diff for concrete cleanup findings only. Do not modify files.", skill: "deslop" },
-      { key: "accessibility", agent: "reviewer", task: "Apply the available 'accessibility' skill to review the UI changes for concrete issues only. Do not modify files.", skill: "accessibility" }
-    ]);
-    return results.map(result => result.output);
-  `,
-  context: "fresh"
-})
+subagent({ workflow: true, context: "fresh" })
 ```
 
 ### Review-loop technique
 
-Use this when the user wants implementation or current diff review to continue until reviewers stop finding fixes worth doing now. Keep the loop in the parent session: one async `worker` implements or fixes, fresh-context `reviewer` agents inspect the actual repo and diff, the parent synthesizes accepted fixes, and one async `worker` applies them (pass `context: "fork"` when that fix pass should inherit the parent thread). The parent can express the sequence up front as an async/background `workflowScript` when the workflow is known, or continue with explicit follow-up workflowScript runs after each async completion. For an initial workflow, pass `async: true` so the main chat is unblocked. Treat an async implementation worker handoff as an intermediate state, not final completion, unless the user explicitly asked for worker-only work, review-only output, or to stop after implementation. Stop when reviewers find no P0 blockers or P1 fixes worth doing now, remaining P2 feedback is optional or deferred, an unapproved product/scope/architecture decision appears, or the max review-round cap is reached. Default to 3 review rounds unless the user sets a different cap. Do not loop for optional polish, and do not let children launch subagents or decide the loop outcome.
+Use this when the user wants implementation or current diff review to continue until reviewers stop finding fixes worth doing now. Keep the loop in the parent session: one async `worker` implements or fixes, fresh-context `reviewer` agents inspect the actual repo and diff, the parent synthesizes accepted fixes, and one async forked `worker` applies them. The parent can express the sequence up front as an async/background workflow script when the workflow is known, or continue with explicit follow-up workflow script runs after each async completion. For an initial workflow, pass `async: true` so the main chat is unblocked. Treat an async implementation worker handoff as an intermediate state, not final completion, unless the user explicitly asked for worker-only work, review-only output, or to stop after implementation. Stop when reviewers find no P0 blockers or P1 fixes worth doing now, remaining P2 feedback is optional or deferred, an unapproved product/scope/architecture decision appears, or the max review-round cap is reached. Default to 3 review rounds unless the user sets a different cap. Do not loop for optional polish, and do not let children launch subagents or decide the loop outcome.
 
-As a conservative orchestration policy, do not pass `turnBudget` or a hard `toolBudget` to an implementation worker, fix worker, reviewer with edit authority, or other mutation-capable child. The default tool budget blocks read/search tools rather than mutation tools, but count limits still do not measure delivery safety. Use a narrow task plus an outer elapsed deadline with enough margin, then request a checkpoint after the current tool returns. The checkpoint should report changed files, build/test state, remaining work, and commit or PR state. An elapsed timeout is not a mutation-safe boundary and must not be used as the checkpoint trigger.
+As a conservative orchestration policy, do not pass a hard `toolBudget` to an implementation worker, fix worker, reviewer with edit authority, or other mutation-capable child. The default tool budget blocks read/search tools rather than mutation tools, but count limits still do not measure delivery safety. Use a narrow task plus an outer elapsed deadline with enough margin, then request a checkpoint after the current tool returns. The checkpoint should report changed files, build/test state, remaining work, and commit or PR state. An elapsed timeout is not a mutation-safe boundary and must not be used as the checkpoint trigger.
 
 ### Parallel research technique
 
@@ -100,59 +109,64 @@ Use this when the question needs both external evidence and local implications.
 
 ### Gather-context-and-clarify technique
 
-Use this at the start of non-trivial work. Launch `scout` for local context and `researcher` only when external docs, recent sources, ecosystem context, or primary evidence would materially improve understanding. Ask children for concise findings plus remaining clarification questions. Then synthesize what is known and use `interview` to ask the unresolved questions needed for shared understanding before planning or implementing.
+Use this when the operator requests delegated context gathering. Launch `scout` for local context and `researcher` only when external docs, recent sources, ecosystem context, or primary evidence would materially improve understanding. Ask children for concise findings plus remaining clarification questions. Then synthesize what is known and use `interview` to ask the unresolved questions needed for shared understanding before planning or implementing.
 
 ### Parallel cleanup technique
 
-Use this after implementation when the user wants cleanup review or when a final pass would reduce AI-slop. Launch two fresh-context `reviewer` tasks with `output: false` and `progress: false`: one deslop pass and one verbosity pass. If the `deslop` or `verbosity-cleaner` skills are available, pass the relevant skill to that reviewer; otherwise inline the criteria. Both reviewers are review-only and should flag concrete issues with severity, file/line references, and smallest safe fixes. Phrase the constraint as “Do not modify project/source files; returning findings through the configured output artifact is allowed” when you use `output` or `outputMode: "file-only"`. The parent decides what to apply and asks before making changes unless cleanup was already authorized.
+Use this after implementation when the user or applicable instructions request delegated cleanup review. Launch two fresh-context `reviewer` tasks with `output: false` and `progress: false`: one deslop pass and one verbosity pass. If the `deslop` or `verbosity-cleaner` skills are available, pass the relevant skill to that reviewer; otherwise inline the criteria. Both reviewers are review-only and should flag concrete issues with severity, file/line references, and smallest safe fixes. Phrase the constraint as “Do not modify project/source files; returning findings through the configured output artifact is allowed” when you use `output` or `outputMode: "file-only"`. The parent decides what to apply and asks before making changes unless cleanup was already authorized.
 
 ### Staged fix orchestration technique
 
-Use this when a broad diff has known reviewer findings across several items and the user wants the parent to “orchestrate subagents like a boss.” Keep the active worktree safe with a three-stage `workflowScript`:
+Use this when a broad diff has known reviewer findings across several items and the user wants the parent to “orchestrate subagents like a boss.” Keep mutation ownership explicit in the workflow script:
 
-1. A parallel read-only planning fanout, one reviewer per issue cluster. Each child inspects the real diff and returns exact files, line refs, proposed fixes, and focused validation. They must not edit.
-2. One writer worker. It receives the reviewer summaries as the awaited planning results (or their durable output paths) interpolated into its task, plus the parent’s accepted scope, stop rules, and verification contract. It is the only child allowed to edit the active worktree.
-3. A parallel read-only validation fanout. Validators inspect the worker diff from fresh context with distinct angles, report pass/fail, remaining blockers, and missing verification.
+When staged seams are available, a low-tier writer should not receive the
+end-to-end issue. Use `runs.lanes` inside the workflow script to keep stages narrow:
+a scout/red test, helper-only change, one render seam, validation, minimality
+challenge, or fresh review. Give the writer only its assigned implementation
+stage; keep sequencing and synthesis with the parent.
+
+1. A parallel read-only planning fanout, one reviewer per issue cluster or review angle. Each child inspects the real diff and returns exact files, line refs, proposed fixes, and focused validation. They must not edit.
+2. Classify the accepted fixes on the lane board before writer launch. If issue clusters are independently testable source or contract boundaries, use isolated component writers with exclusive file/contract ownership and focused gates. Each component must commit or produce another durable handoff; only after those handoffs are ready, start one integration-only owner.
+3. Use one writer for the complete fix only when the accepted fixes are one tightly coupled existing diff/seam and the lane board records evidence that splitting would create overlapping ownership or artificial handoffs. Give that writer the planning summaries, accepted scope, stop rules, and verification contract.
+4. Run fresh-context, read-only validation against the integrated or single-seam result. Validators use distinct angles and report pass/fail, remaining blockers, and missing verification.
 
 Prefer `async: true`, `context: "fresh"` for reviewers/validators, `outputMode: "file-only"` for large summaries, and per-stage output names that will not collide. Use stable `runs` keys plus `phase` and `label` on each launch item to make async status readable, and hold each awaited result in an ordinary JavaScript variable when a later step needs that specific result — interpolate it (or the durable output path you declared for that child) into the later task text instead of passing a whole aggregate blob. Use this pattern instead of launching several writer workers into a dirty worktree. Include non-blocking suggestions in the writer prompt only when they are small, safe, and do not expand product scope; otherwise record them as deferred.
 
 When one child returns a structured target list, use ordinary JavaScript to validate/filter it and map bounded entries into `runs.all`; do not use the removed chain fanout DSL.
 
-Example shape:
+Single-seam example shape (the lane board records that these are review angles on one tightly coupled deployment-lifecycle diff, not independent implementation contracts):
+
+```js workflow
+// Stage 1: parallel read-only review angles on one recorded single seam
+const plans = await runs.all([
+  { key: "correctness-plan", agent: "reviewer", phase: "Planning", label: "Review lifecycle correctness", task: "Plan correctness fixes for the same deployment-lifecycle diff. Do not modify project/source files; returning findings via the configured output artifact is allowed.", output: "plans/correctness.md", outputMode: "file-only" },
+  { key: "regression-plan", agent: "reviewer", phase: "Planning", label: "Review lifecycle regressions", task: "Plan regression coverage for the same deployment-lifecycle diff. Do not modify project/source files; returning findings via the configured output artifact is allowed.", output: "plans/regressions.md", outputMode: "file-only" },
+  { key: "minimality-plan", agent: "reviewer", phase: "Planning", label: "Review lifecycle minimality", task: "Plan minimality fixes for the same deployment-lifecycle diff. Do not modify project/source files; returning findings via the configured output artifact is allowed.", output: "plans/minimality.md", outputMode: "file-only" }
+]);
+
+// Stage 2: one writer for this lane-board-recorded single seam.
+// Under outputMode "file-only" the awaited .output is the saved-output
+// reference, so pass those managed artifact references to the writer.
+const worker = await runs.run("apply-fixes", {
+  agent: "worker",
+  phase: "Implementation",
+  label: "Apply accepted fixes",
+  task: "Apply only the accepted fixes to the one deployment-lifecycle seam. The lane board records that splitting this existing diff would create overlapping ownership. You are the sole writer for this seam. Run focused validation and report changed files, commands, failures, and remaining issues.\n\nCorrectness plan: " + plans[0].output + "\n\nRegression plan: " + plans[1].output + "\n\nMinimality plan: " + plans[2].output,
+  output: "worker/fixes.md",
+  outputMode: "file-only"
+});
+
+// Stage 3: parallel read-only validation fanout
+const validations = await runs.all([
+  { key: "validate-behavior", agent: "reviewer", phase: "Validation", label: "Validate lifecycle behavior", task: "Validate behavior in the post-worker deployment-lifecycle diff. Start from the worker result: " + worker.output + ". Do not modify project/source files; returning findings via the configured output artifact is allowed.", output: "validation/behavior.md", outputMode: "file-only" },
+  { key: "validate-minimality", agent: "reviewer", phase: "Validation", label: "Validate lifecycle minimality", task: "Validate minimality in the same post-worker deployment-lifecycle diff. Start from the worker result: " + worker.output + ". Do not modify project/source files; returning findings via the configured output artifact is allowed.", output: "validation/minimality.md", outputMode: "file-only" }
+]);
+
+return { worker: worker.output, validations: validations.map(v => v.output) };
+```
 
 ```typescript
-subagent({
-  async: true,
-  context: "fresh",
-  workflowScript: `
-    // Stage 1: parallel read-only planning fanout (stable keys, one per issue cluster)
-    const plans = await runs.all([
-      { key: "deploy-plan", agent: "reviewer", phase: "Planning", label: "Deploy docs", task: "Plan fixes for deploy docs/workflow. Inspect the current diff. Do not modify project/source files; returning findings via the configured output artifact is allowed.", output: "plans/deploy.md", outputMode: "file-only" },
-      { key: "scheduler-plan", agent: "reviewer", phase: "Planning", label: "Scheduler contract", task: "Plan fixes for scheduler contract. Inspect the current diff. Do not modify project/source files; returning findings via the configured output artifact is allowed.", output: "plans/scheduler.md", outputMode: "file-only" },
-      { key: "sandbox-plan", agent: "reviewer", phase: "Planning", label: "Sandbox/security", task: "Plan fixes for sandbox/security. Inspect the current diff. Do not modify project/source files; returning findings via the configured output artifact is allowed.", output: "plans/sandbox.md", outputMode: "file-only" }
-    ]);
-
-    // Stage 2: single writer — the only child allowed to edit the active worktree.
-    // Under outputMode "file-only" the awaited .output is the saved-output
-    // reference, so pass those managed artifact references to the writer.
-    const worker = await runs.run("apply-fixes", {
-      agent: "worker",
-      phase: "Implementation",
-      label: "Apply accepted fixes",
-      task: "Apply only the accepted fixes from these planning summaries. You are the sole writer for the active worktree. Run focused validation and report changed files, commands, failures, and remaining issues.\\n\\nDeploy plan: " + plans[0].output + "\\n\\nScheduler plan: " + plans[1].output + "\\n\\nSandbox plan: " + plans[2].output,
-      output: "worker/fixes.md",
-      outputMode: "file-only"
-    });
-
-    // Stage 3: parallel read-only validation fanout
-    const validations = await runs.all([
-      { key: "validate-deploy-scheduler", agent: "reviewer", phase: "Validation", label: "Deploy/scheduler validation", task: "Validate the post-worker diff for deploy and scheduler fixes. Start from the worker result: " + worker.output + ". Do not modify project/source files; returning findings via the configured output artifact is allowed.", output: "validation/deploy-scheduler.md", outputMode: "file-only" },
-      { key: "validate-sandbox", agent: "reviewer", phase: "Validation", label: "Sandbox validation", task: "Validate the post-worker diff for sandbox/security fixes. Start from the worker result: " + worker.output + ". Do not modify project/source files; returning findings via the configured output artifact is allowed.", output: "validation/sandbox.md", outputMode: "file-only" }
-    ]);
-
-    return { worker: worker.output, validations: validations.map(v => v.output) };
-  `
-})
+subagent({ workflow: true, async: true, context: "fresh" })
 ```
 
 ## Builtin Agents
@@ -160,21 +174,21 @@ subagent({
 Builtin agents load at the lowest priority. Project agents override user agents,
 and user/project agents override builtins with the same name.
 
-| Agent | Purpose | Model | Typical output / role |
+| Agent | Purpose | Recommended tier | Typical output / role |
 |-------|---------|-------|------------------------|
-| `scout` | Fast codebase recon | inherits default | Writes `context.md` handoff material |
-| `worker` | Implementation and approved oracle handoffs | inherits default | Single-writer implementation with decision escalation |
-| `reviewer` | Review specialist | inherits default | Default recipes are review-only; tools include edit/write when a fix pass is explicit |
-| `researcher` | Web research brief generator | inherits default | Writes `research.md` |
-| `delegate` | Lightweight generic delegate | inherits default | No fixed output; generic delegated work |
-| `oracle` | Decision-consistency advisory review | inherits default | Advisory review, intercom coordination |
-| `advisor` | Claude Code-compatible alias for `oracle` | inherits default | Same advisory role as `oracle` |
+| `scout` | Fast codebase recon | fast worker/scout tier | Writes `context.md` handoff material |
+| `worker` | Implementation and approved oracle handoffs | capable worker tier | Single-writer implementation with decision escalation |
+| `reviewer` | Review specialist | strong reviewer tier; high thinking for serious reviews | Default recipes are review-only; tools include edit/write when a fix pass is explicit |
+| `researcher` | Web research brief generator | inherits configured default | Writes `research.md` |
+| `delegate` | Lightweight generic delegate | inherits configured default | No fixed output; generic delegated work |
+| `oracle` | Rare hard-decision/root-cause escalation | top-reasoning critic tier, bounded read-only; high thinking escalation only | Advisory trajectory review, not routine code review |
+| `advisor` | Compatibility alias for `oracle` | top-reasoning critic tier, bounded read-only; high thinking escalation only | Same advisory escalation role as `oracle` |
 
-Builtin `worker` and `delegate` use strict tool allowlists and do not inherit ambient parent extension tools. To give a child an extension tool, name it in `tools` and load its provider via `extensions`, a path-like `tools` entry, or `subagentOnlyExtensions`. Custom agents without an `extensions` field follow `subagents.defaultExtensions` when set.
+Builtin `worker` and `delegate` use strict tool allowlists and do not inherit ambient parent extension tools. To give a child an extension tool, name it in `tools` and load its provider via `extensions`, a path-like `tools` entry, or `subagentOnlyExtensions`. Agents without the corresponding field follow `subagents.defaultExtensions` or the ambient-preserving `subagents.defaultSubagentOnlyExtensions` when set.
 
-Builtin agents inherit the current Pi default model unless a run, user setting, project setting, or `subagents.defaultModel` overrides `model`. Set `subagents.defaultModel` when subagents should use a different default model than the parent session. Override builtin defaults before copying full agent files when a small tweak is enough.
+Builtin agents inherit the current Pi default model unless a run, user setting, project setting, or `subagents.defaultModel` overrides `model`. The table records recommended tier routing, not shipped hard defaults; explicit run, user, or project settings still win. Keep the parent/orchestrator on the ordinary strong default model unless parent/user policy says otherwise. Override builtin defaults before copying full agent files when a small tweak is enough.
 
-Set `subagents.defaultThinking` to apply a shared thinking level to builtin, package, user, and project agents whose frontmatter leaves `thinking` unset. Project settings win over user settings; explicit frontmatter (including `thinking: false`), `agentOverrides.<name>.thinking`, and per-run overrides remain more specific. This setting affects child agents only and does not change the parent session's default thinking level.
+Set `subagents.defaultThinking` to apply a shared thinking level to builtin, package, user, and project agents whose frontmatter leaves `thinking` unset. Project settings win over user settings; matching `agentOverrides.<name>.thinking` and per-run overrides replace frontmatter, while an explicit frontmatter value remains in effect when no matching override is set. This setting affects child agents only and does not change the parent session's default thinking level.
 
 ```json
 {
@@ -187,12 +201,32 @@ Set `subagents.defaultThinking` to apply a shared thinking level to builtin, pac
 For one run, use inline config:
 
 ```text
-/run reviewer[model=anthropic/claude-sonnet-4] "Review this diff"
+/run reviewer[model=provider/review-model] "Review this diff"
 ```
 
 For persistent tweaks, edit `subagents.agentOverrides` in user or project settings. User overrides apply everywhere. Project overrides apply only in that repo and win over user overrides. Use `/subagents-models` or `subagent({ action: "models" })` to inspect the live mapping after settings and overrides load.
 
-Model ids do not have to be exact. Separator variations (`claude-haiku-4.5` vs `claude-haiku-4-5`), case (`Claude-Sonnet-4`), and optional trailing date stamps (`claude-haiku-4-5-20251001`) all resolve to the same registry model. Exact `provider/id` wins; a qualified `provider/model` never switches providers. To constrain subagents to a budget or compliance profile, set `subagents.modelScope: { enforce: true, allow: ["anthropic/*", "openai/gpt-5-*"] }` in user or project settings. Out-of-scope models you pass explicitly error and abort; models inherited from frontmatter, `subagents.defaultModel`, agent frontmatter, or the parent session only warn.
+Provider-scoped entries can layer on top of the default override for the active parent session provider. The provider is selected once from the parent model. Within each settings file, the provider entry wins per field; project settings still win over user settings.
+
+```json
+{
+  "subagents": {
+    "agentOverrides": {
+      "worker": { "thinking": "medium" }
+    },
+    "agentOverridesByProvider": {
+      "provider-a": {
+        "worker": { "model": "provider-a/fast-worker-model" }
+      },
+      "provider-b": {
+        "worker": { "model": "provider-b/fast-worker-model" }
+      }
+    }
+  }
+}
+```
+
+Model ids do not have to be exact. Separator variations (`fast.worker-v1` vs `fast-worker-v1`), case (`Strong-Review-Model`), and optional trailing date stamps all resolve to the same registry model. Exact `provider/id` wins; a qualified `provider/model` never switches providers. To constrain subagents to a budget or compliance profile, set `subagents.modelScope: { enforce: true, allow: ["approved-provider/*", "second-provider/approved-*"] }` in user or project settings. Out-of-scope models you pass explicitly error and abort; models inherited from frontmatter, `subagents.defaultModel`, agent frontmatter, or the parent session only warn.
 
 For model fleets, use the profile commands instead of hand-editing repeated overrides: `/subagents-refresh-provider-models <provider>`, `/subagents-generate-profiles <provider>`, `/subagents-load-profile <name>`, and `/subagents-check-profile <name>`. Profiles live under `~/.selesai/agent/profiles/pi-subagents/` and replace only `settings.subagents` when loaded.
 
@@ -206,7 +240,7 @@ A strong subagent prompt usually includes:
 - **Authority boundary**: whether the child may read, edit, commit, push, comment, close, merge, publish, or release. Omit or forbid actions that are not approved.
 - **Context/evidence**: relevant plan paths, files, diffs, decisions, or user constraints already approved.
 - **Success criteria**: what must be true before the child can finish.
-- **Hard constraints**: true invariants only, such as no edits for review-only tasks, one writer thread, child must not run subagents unless it is an explicitly assigned `tools: subagent` fanout child, or escalation for unapproved decisions.
+- **Hard constraints**: true invariants only, such as no edits for review-only tasks, one writer thread, child must not run subagents unless it is explicitly authorized through `tools: subagent` or `allowNestedSubagents: true`, or escalation for unapproved decisions.
 - **Validation**: targeted checks to run, or the next-best check when validation is impossible.
 - **Output**: the expected summary shape, artifact path, or finding format. Use managed artifact paths for scratch reports; reserve repo-qualified absolute paths for durable handoffs that the user approved.
 - **Stop rules**: when to ask via `intercom` or `contact_supervisor`, when to stop after enough evidence, and when not to keep searching.
@@ -228,9 +262,8 @@ Direct settings example:
   "subagents": {
     "agentOverrides": {
       "reviewer": {
-        "model": "anthropic/claude-sonnet-4",
+        "model": "provider/strong-review-model",
         "thinking": "high",
-        "fallbackModels": ["openai/gpt-5-mini"],
         "acceptanceRole": "read-only"
       }
     }
@@ -238,8 +271,8 @@ Direct settings example:
 }
 ```
 
-Useful override fields: `description`, `model`, `fallbackModels`, `thinking`,
-`systemPromptMode`, `inheritProjectContext`, `inheritSkills`, `defaultContext`,
+Useful override fields: `description`, `model`, `thinking`,
+`systemPromptMode`, `inheritProjectContext`, `inheritGlobalContext`, `inheritSkills`, `defaultContext`,
 `acceptanceRole`, `disabled`, `skills`, `tools`, `extensions`, and `systemPrompt`.
 `description` replaces the discovered description for builtin and custom agents
 in `list` output, which is useful for deployment-specific routing notes.
@@ -248,20 +281,17 @@ agent with the same name only when you want a substantially different agent.
 
 ### Recommended model tiering (optional)
 
-When several providers are available, route agents by task shape instead of one model for everything:
+Keep the parent/orchestrator on the ordinary strong default model because omission failures are cheaper than unnecessary commissions. Route workers and scouts to a fast, capable worker tier, and keep serious reviews on the strong reviewer tier at high thinking. Do not use `oracle` or a top-reasoning model as the routine fresh-review default. Use that tier only for bounded, read-only critic/oracle/root-cause audits after ordinary review, CI, bot, or source evidence is insufficient; critic-tier high thinking is escalation-only and never an autonomous root. Explicit parent/user model policy wins over these recommendations.
 
-1. **Fast workhorse** — cheapest capable model at low thinking for recon, lookups, and mechanical edits (for example on `scout`).
-2. **Standard well-scoped** — mid-tier model at medium thinking for most delegations: routine multi-file edits, focused reviews, straightforward implementation (for example on `worker`, `reviewer`, `delegate`).
-3. **Deep but bounded** — top reasoning model at high thinking only for hard tasks that arrive with explicit goals and completion criteria; these models loop on vague goals (for example on oracle-style agents).
-4. **Taste and intent** — a model that reads human intent well for ambiguous work: UX/design judgment, product tradeoffs, planning from vague requirements, writing quality.
+Examples are illustrative, not requirements. Map these tiers to concrete models in user/project settings or a profile. A non-OpenAI setup should choose comparable available models by capability.
 
-Routing rule: use tiers 1–3 when the task is well-scoped; use tier 4 when scoping or judging is the task itself. Give tier-4 agents cross-provider `fallbackModels` so subscription usage limits degrade gracefully; fallback triggers automatically on rate-limit and overload errors. Note that forked context over an Anthropic parent transcript with signed thinking blocks forces the child's thinking off, so intent-tier agents work best with fresh context.
+Each child launch uses one resolved model exactly once. If quota or availability fails, surface that failure and let the parent or operator explicitly launch a later attempt with another model. Forked children keep their requested thinking level even when provider-specific reasoning blocks are stripped from the inherited transcript.
 
 If a provider rejects model IDs with thinking suffixes, use
 `subagents.disableThinking: true` in user or project settings to clear bundled
 builtin thinking defaults globally. A higher-precedence per-agent `thinking`
-override can opt one builtin back in. Existing custom-agent frontmatter remains authoritative.
+override can opt one builtin back in or replace custom-agent frontmatter thinking.
 
-Set `subagents.defaultExtensions` to give agents without an `extensions` field a shared child extension allowlist. Omit it to preserve ambient extension discovery, set it to `[]` to disable ambient extensions by default, or use `agentOverrides.<name>.extensions` for one agent. Explicit custom-agent frontmatter still wins.
+Set `subagents.defaultExtensions` to give agents without an `extensions` field a shared child extension allowlist. Omit it to preserve ambient extension discovery, set it to `[]` to disable ambient extensions by default, or use `agentOverrides.<name>.extensions` for one agent. Set `subagents.defaultSubagentOnlyExtensions` to add shared child-only paths without disabling ambient discovery. For either field, explicit frontmatter suppresses the default and a matching override replaces or false-clears it; lists are not combined.
 
-Tool description modes live in `~/.selesai/agent/extensions/subagent/config.json`, not `subagents` settings. The default uses split prompt metadata: a short tool description plus active `promptSnippet` and `promptGuidelines`. Set `toolDescriptionMode` to `full` or `compact` to force one description string, or `custom` to read `subagent-tool-description.md` from the project config dir or agent dir; invalid custom files fall back to full mode and the safety guidance is still appended.
+Tool description modes live in `~/.selesai/agent/extensions/subagent/config.json`, not `subagents` settings. The default is `full`: one complete tool description. Set `toolDescriptionMode` to `compact` for a shorter description, or `custom` to read `subagent-tool-description.md` from the project config dir or agent dir; invalid custom files fall back to full mode and the safety guidance is still appended.

package/dist/skills/pi-subagents/references/review-and-validation.md

@@ -0,0 +1,73 @@
+# Pi Subagents: Review And Validation
+
+Generic review and delivery guidance for operator-authorized delegated work. This file does not encode private backlog, merge, or release policy.
+
+## Delivery loop
+
+Use the smallest loop that proves the change:
+
+1. Inspect the source, diff, issue, or plan directly.
+2. Keep one writer for each cwd or worktree.
+3. Run focused validation that can fail for the changed behavior.
+4. When the operator/project delegation contract calls for independent review, use a fresh-context read-only reviewer; otherwise parent inspection is valid.
+5. Apply only accepted findings inside the same writer boundary.
+6. Re-run affected validation and review only the changed blast radius.
+7. Inspect the final diff and evidence before parent acceptance.
+
+Skip review ceremony for trivial wording, renames, or local-only probes when direct parent inspection is enough.
+
+## Review shape
+
+| Situation | Shape |
+| --- | --- |
+| One coherent diff or one risk | one reviewer |
+| Independent risks, such as correctness, tests, security, or UI | parallel reviewers with distinct contracts |
+| Possible over-scope or needless complexity | same-writer challenge before fresh review |
+| Material design tradeoff | council mode |
+
+Reviewers are fresh-context by default. Use the ordinary `reviewer` role for routine code review. Forked oracle/advisor runs are escalation-only for parent-history, drift, root-cause, model-routing, or hard tradeoff evidence.
+
+## Finding disposition
+
+The parent classifies each finding against current HEAD:
+
+- **Valid blocker:** concrete failure, repro, security issue, contract mismatch, or source-proven regression. Fix now.
+- **Valid non-blocker:** real but outside the delivery slice. Record or defer.
+- **Stale:** fixed or absent at the reviewed head. Cite current evidence.
+- **Invalid:** contradicted by source, tests, docs, or user-approved scope. Cite the contradiction.
+- **Out of policy/scope:** needs unapproved product, architecture, authority, release, or public-repo action. Escalate.
+- **Speculative:** no contract, repro, or reachable failure. Do not block.
+
+A clean reviewer result is evidence, not publication authority.
+
+## Gate-failure triage
+
+When validation fails:
+
+1. Confirm the run belongs to the exact head/ref under judgment.
+2. Read the focused failing logs first.
+3. Name the failing test, assertion, contract, or thread.
+4. Classify cause: current diff, stale test, environment/setup, or existing flake.
+5. Reproduce locally when practical with the narrowest command.
+6. Patch forward when the current diff caused it.
+7. For stale/flaky failures, collect proof before one rerun or residual-risk note.
+8. Re-run the affected command or exact-head gate after every fix.
+
+For bot comments, classify each thread as valid, stale, invalid, or out of policy before assigning severity.
+
+## Final checklist
+
+Before reporting delegated work as done, verify the relevant subset:
+
+- final diff contains only intended files
+- focused validation covers changed behavior
+- required independent review has fresh-review evidence
+- accepted findings are fixed and revalidated
+- publication authority exists before push, comment, close, merge, deploy, or release
+- external checks are exact-head when used as evidence
+- handoff is durable before cleanup
+- residual risks, skipped validation, and blocked decisions are explicit
+
+## Public/private boundary
+
+For issue/PR backlogs, releases, merge queues, contributor credit, or repo-specific policy, load the matching user/project skill when available. Keep those rules out of this public package until intentionally released.

package/dist/utils/ansi.d.ts

@@ -1 +1,9 @@
+/**
+ * Split streamed text into a part that is safe to pass to stripAnsi now and a trailing
+ * unfinished escape sequence that should be prepended to the next chunk.
+ */
+export declare function splitIncompleteAnsiSuffix(value: string): {
+    complete: string;
+    pending: string;
+};
 export declare function stripAnsi(value: string): string;

package/dist/utils/ansi.js

@@ -25,17 +25,36 @@
  * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
  * SOFTWARE.
  */
-function ansiRegex({ onlyFirst = false } = {}) {
-    // Valid string terminator sequences are BEL, ESC\, and 0x9c
-    const ST = "(?:\\u0007|\\u001B\\u005C|\\u009C)";
-    // OSC sequences only: ESC ] ... ST (non-greedy until the first ST)
-    const osc = `(?:\\u001B\\][\\s\\S]*?${ST})`;
-    // CSI and related: ESC/C1, optional intermediates, optional params (supports ; and :) then final byte
-    const csi = "[\\u001B\\u009B][[\\]()#;?]*(?:\\d{1,4}(?:[;:]\\d{0,4})*)?[\\dA-PR-TZcf-nq-uy=><~]";
-    const pattern = `${osc}|${csi}`;
-    return new RegExp(pattern, onlyFirst ? undefined : "g");
+// Valid string terminator sequences are BEL, ESC\, and 0x9c
+const ST = "(?:\\u0007|\\u001B\\u005C|\\u009C)";
+// OSC sequences: ESC ] ... ST
+const OSC_START = "\\u001B\\]";
+// CSI and related: ESC/C1, optional intermediates, optional params (supports ; and :), then final byte
+const CSI_START = "[\\u001B\\u009B][[\\]()#;?]*(?:\\d{1,4}(?:[;:]\\d{0,4})*)?";
+const CSI_FINAL = "[\\dA-PR-TZcf-nq-uy=><~]";
+// Complete sequences. OSC is non-greedy until the first ST.
+const ansiRegex = new RegExp(`(?:${OSC_START}[\\s\\S]*?${ST})|${CSI_START}${CSI_FINAL}`, "g");
+// Unfinished sequence at the end of the text: OSC without its ST (a trailing ESC may start ESC\),
+// or CSI without its final byte.
+const unfinishedAnsiAtEndRegex = new RegExp(`(?:${OSC_START}(?:[^\\u0007\\u009C\\u001B]|\\u001B(?!\\\\))*|${CSI_START})$`);
+// Longest unfinished sequence held back while streaming. Longer ones are processed as-is.
+const MAX_PENDING_ANSI_LENGTH = 256;
+/**
+ * Split streamed text into a part that is safe to pass to stripAnsi now and a trailing
+ * unfinished escape sequence that should be prepended to the next chunk.
+ */
+export function splitIncompleteAnsiSuffix(value) {
+    if (!value.includes("\u001B") && !value.includes("\u009B")) {
+        return { complete: value, pending: "" };
+    }
+    const windowStart = Math.max(0, value.length - MAX_PENDING_ANSI_LENGTH);
+    const match = unfinishedAnsiAtEndRegex.exec(value.slice(windowStart));
+    if (!match) {
+        return { complete: value, pending: "" };
+    }
+    const splitAt = windowStart + match.index;
+    return { complete: value.slice(0, splitAt), pending: value.slice(splitAt) };
 }
-const regex = ansiRegex();
 export function stripAnsi(value) {
     if (typeof value !== "string") {
         throw new TypeError(`Expected a \`string\`, got \`${typeof value}\``);
@@ -47,5 +66,5 @@ export function stripAnsi(value) {
     // Even though the regex is global, we don't need to reset the `.lastIndex`
     // because unlike `.exec()` and `.test()`, `.replace()` does it automatically
     // and doing it manually has a performance penalty.
-    return value.replace(regex, "");
+    return value.replace(ansiRegex, "");
 }