ytdb-slate 0.1.0

package/LICENSE

@@ -0,0 +1,201 @@
+                                 Apache License
+                           Version 2.0, January 2004
+                        http://www.apache.org/licenses/
+
+   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+   1. Definitions.
+
+      "License" shall mean the terms and conditions for use, reproduction,
+      and distribution as defined by Sections 1 through 9 of this document.
+
+      "Licensor" shall mean the copyright owner or entity authorized by
+      the copyright owner that is granting the License.
+
+      "Legal Entity" shall mean the union of the acting entity and all
+      other entities that control, are controlled by, or are under common
+      control with that entity. For the purposes of this definition,
+      "control" means (i) the power, direct or indirect, to cause the
+      direction or management of such entity, whether by contract or
+      otherwise, or (ii) ownership of fifty percent (50%) or more of the
+      outstanding shares, or (iii) beneficial ownership of such entity.
+
+      "You" (or "Your") shall mean an individual or Legal Entity
+      exercising permissions granted by this License.
+
+      "Source" form shall mean the preferred form for making modifications,
+      including but not limited to software source code, documentation
+      source, and configuration files.
+
+      "Object" form shall mean any form resulting from mechanical
+      transformation or translation of a Source form, including but
+      not limited to compiled object code, generated documentation,
+      and conversions to other media types.
+
+      "Work" shall mean the work of authorship, whether in Source or
+      Object form, made available under the License, as indicated by a
+      copyright notice that is included in or attached to the work
+      (an example is provided in the Appendix below).
+
+      "Derivative Works" shall mean any work, whether in Source or Object
+      form, that is based on (or derived from) the Work and for which the
+      editorial revisions, annotations, elaborations, or other modifications
+      represent, as a whole, an original work of authorship. For the purposes
+      of this License, Derivative Works shall not include works that remain
+      separable from, or merely link (or bind by name) to the interfaces of,
+      the Work and Derivative Works thereof.
+
+      "Contribution" shall mean any work of authorship, including
+      the original version of the Work and any modifications or additions
+      to that Work or Derivative Works thereof, that is intentionally
+      submitted to Licensor for inclusion in the Work by the copyright owner
+      or by an individual or Legal Entity authorized to submit on behalf of
+      the copyright owner. For the purposes of this definition, "submitted"
+      means any form of electronic, verbal, or written communication sent
+      to the Licensor or its representatives, including but not limited to
+      communication on electronic mailing lists, source code control systems,
+      and issue tracking systems that are managed by, or on behalf of, the
+      Licensor for the purpose of discussing and improving the Work, but
+      excluding communication that is conspicuously marked or otherwise
+      designated in writing by the copyright owner as "Not a Contribution."
+
+      "Contributor" shall mean Licensor and any individual or Legal Entity
+      on behalf of whom a Contribution has been received by Licensor and
+      subsequently incorporated within the Work.
+
+   2. Grant of Copyright License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      copyright license to reproduce, prepare Derivative Works of,
+      publicly display, publicly perform, sublicense, and distribute the
+      Work and such Derivative Works in Source or Object form.
+
+   3. Grant of Patent License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      (except as stated in this section) patent license to make, have made,
+      use, offer to sell, sell, import, and otherwise transfer the Work,
+      where such license applies only to those patent claims licensable
+      by such Contributor that are necessarily infringed by their
+      Contribution(s) alone or by combination of their Contribution(s)
+      with the Work to which such Contribution(s) was submitted. If You
+      institute patent litigation against any entity (including a
+      cross-claim or counterclaim in a lawsuit) alleging that the Work
+      or a Contribution incorporated within the Work constitutes direct
+      or contributory patent infringement, then any patent licenses
+      granted to You under this License for that Work shall terminate
+      as of the date such litigation is filed.
+
+   4. Redistribution. You may reproduce and distribute copies of the
+      Work or Derivative Works thereof in any medium, with or without
+      modifications, and in Source or Object form, provided that You
+      meet the following conditions:
+
+      (a) You must give any other recipients of the Work or
+          Derivative Works a copy of this License; and
+
+      (b) You must cause any modified files to carry prominent notices
+          stating that You changed the files; and
+
+      (c) You must retain, in the Source form of any Derivative Works
+          that You distribute, all copyright, patent, trademark, and
+          attribution notices from the Source form of the Work,
+          excluding those notices that do not pertain to any part of
+          the Derivative Works; and
+
+      (d) If the Work includes a "NOTICE" text file as part of its
+          distribution, then any Derivative Works that You distribute must
+          include a readable copy of the attribution notices contained
+          within such NOTICE file, excluding those notices that do not
+          pertain to any part of the Derivative Works, in at least one
+          of the following places: within a NOTICE text file distributed
+          as part of the Derivative Works; within the Source form or
+          documentation, if provided along with the Derivative Works; or,
+          within a display generated by the Derivative Works, if and
+          wherever such third-party notices normally appear. The contents
+          of the NOTICE file are for informational purposes only and
+          do not modify the License. You may add Your own attribution
+          notices within Derivative Works that You distribute, alongside
+          or as an addendum to the NOTICE text from the Work, provided
+          that such additional attribution notices cannot be construed
+          as modifying the License.
+
+      You may add Your own copyright statement to Your modifications and
+      may provide additional or different license terms and conditions
+      for use, reproduction, or distribution of Your modifications, or
+      for any such Derivative Works as a whole, provided Your use,
+      reproduction, and distribution of the Work otherwise complies with
+      the conditions stated in this License.
+
+   5. Submission of Contributions. Unless You explicitly state otherwise,
+      any Contribution intentionally submitted for inclusion in the Work
+      by You to the Licensor shall be under the terms and conditions of
+      this License, without any additional terms or conditions.
+      Notwithstanding the above, nothing herein shall supersede or modify
+      the terms of any separate license agreement you may have executed
+      with Licensor regarding such Contributions.
+
+   6. Trademarks. This License does not grant permission to use the trade
+      names, trademarks, service marks, or product names of the Licensor,
+      except as required for reasonable and customary use in describing the
+      origin of the Work and reproducing the content of the NOTICE file.
+
+   7. Disclaimer of Warranty. Unless required by applicable law or
+      agreed to in writing, Licensor provides the Work (and each
+      Contributor provides its Contributions) on an "AS IS" BASIS,
+      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+      implied, including, without limitation, any warranties or conditions
+      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+      PARTICULAR PURPOSE. You are solely responsible for determining the
+      appropriateness of using or redistributing the Work and assume any
+      risks associated with Your exercise of permissions under this License.
+
+   8. Limitation of Liability. In no event and under no legal theory,
+      whether in tort (including negligence), contract, or otherwise,
+      unless required by applicable law (such as deliberate and grossly
+      negligent acts) or agreed to in writing, shall any Contributor be
+      liable to You for damages, including any direct, indirect, special,
+      incidental, or consequential damages of any character arising as a
+      result of this License or out of the use or inability to use the
+      Work (including but not limited to damages for loss of goodwill,
+      work stoppage, computer failure or malfunction, or any and all
+      other commercial damages or losses), even if such Contributor
+      has been advised of the possibility of such damages.
+
+   9. Accepting Warranty or Additional Liability. While redistributing
+      the Work or Derivative Works thereof, You may choose to offer,
+      and charge a fee for, acceptance of support, warranty, indemnity,
+      or other liability obligations and/or rights consistent with this
+      License. However, in accepting such obligations, You may act only
+      on Your own behalf and on Your sole responsibility, not on behalf
+      of any other Contributor, and only if You agree to indemnify,
+      defend, and hold each Contributor harmless for any liability
+      incurred by, or claims asserted against, such Contributor by reason
+      of your accepting any such warranty or additional liability.
+
+   END OF TERMS AND CONDITIONS
+
+   APPENDIX: How to apply the Apache License to your work.
+
+      To apply the Apache License to your work, attach the following
+      boilerplate notice, with the fields enclosed by brackets "{}"
+      replaced with your own identifying information. (Don't include
+      the brackets!)  The text should be enclosed in the appropriate
+      comment syntax for the file format. We also recommend that a
+      file or class name and description of purpose be included on the
+      same "printed page" as the copyright notice for easier
+      identification within third-party archives.
+
+   Copyright {yyyy} {name of copyright owner}
+
+   Licensed under the Apache License, Version 2.0 (the "License");
+   you may not use this file except in compliance with the License.
+   You may obtain a copy of the License at
+
+       http://www.apache.org/licenses/LICENSE-2.0
+
+   Unless required by applicable law or agreed to in writing, software
+   distributed under the License is distributed on an "AS IS" BASIS,
+   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+   See the License for the specific language governing permissions and
+   limitations under the License.

package/README.md

@@ -0,0 +1,72 @@
+# ytdb-slate
+
+Slate is a thread-weaving orchestration extension for the [pi coding agent](https://github.com/earendil-works/pi-coding-agent).
+
+The orchestrator (your main pi session) dispatches **bounded actions** to persistent **worker threads**. Each completed action is compressed by an LLM into an **episode** — a durable, structured record (intent, actions, findings, artifacts, open issues, handoff notes) that the orchestrator composes into further dispatches instead of re-reading raw transcripts. On top of that, Slate injects a mandatory workflow doctrine: research before design, design review, adversarial review, and track review — with optional umbrella **draft-PR publishing** for tracks.
+
+## Install
+
+Install into a project (pinned, project-scoped — recorded in `.pi/settings.json`):
+
+```bash
+pi install -l npm:ytdb-slate@<version>
+```
+
+> **Note on pinning:** pinned specs (`@<version>`) are deliberately skipped by `pi update --extensions` / `pi update --all`. Bumping the pin is a conscious project change — review Slate's shipped workflow docs for changes when you do.
+
+## Configuration
+
+Optional config file: `slate.json` in the project's pi config dir (`.pi/slate.json`). It is honored **only in trusted projects** (see [Trust](#trust)).
+
+| Key | Type | Default | Semantics |
+| --- | --- | --- | --- |
+| `orchestratorModeDefault` | boolean | `false` | Start fresh interactive sessions with orchestrator mode ON. |
+| `episodeModel` | string | newest available Anthropic Sonnet, else the worker's own model | Model (`provider/id`) used to compress a finished action into an episode. |
+| `workerTools` | string[] | `["read", "bash", "edit", "write", "grep", "find", "ls"]` | Tools available to worker threads (an empty list also falls back to the default). |
+| `maxConcurrent` | number | `4` | Maximum number of worker actions running concurrently. |
+| `pauseThresholdPercent` | number | `40` | Orchestrator context budget (percent, in `(0, 100]`) at which Slate auto-pauses and prepares a fresh-session handoff. |
+| `orchestratorPromptDocs` | string[] | `[]` | Project markdown files (paths relative to the project root) whose **contents** are appended to the orchestrator system prompt. |
+| `workerPromptDocs` | string[] | `[]` | Project markdown files whose **contents** are appended to every worker-thread system prompt. |
+| `workflow.draftPRs` | boolean | `false` | Enable umbrella draft-PR publishing for tracks. |
+| `doctrineExtraPath` | string | — | Project markdown whose **content** is appended to the orchestrator doctrine (project-specific workflow additions). |
+| `reviewPerspectivesPath` | string | — | Project review charters, each declaring its own finding-ID prefix. The doctrine references this **path**; the orchestrator reads the file alongside the shipped review rules. |
+
+Example `.pi/slate.json` (the `docs/agents/...` paths are placeholders — point them at markdown files that actually exist in **your** project):
+
+```json
+{
+  "orchestratorModeDefault": true,
+  "episodeModel": "anthropic/claude-sonnet-5",
+  "maxConcurrent": 4,
+  "pauseThresholdPercent": 40,
+  "orchestratorPromptDocs": ["docs/agents/orchestrator-guidelines.md"],
+  "workerPromptDocs": ["docs/agents/thread-guidelines.md"],
+  "workflow": { "draftPRs": true },
+  "doctrineExtraPath": "docs/agents/workflow-additions.md",
+  "reviewPerspectivesPath": "docs/agents/review-perspectives.md"
+}
+```
+
+> **Silent skip:** the project-file keys fail silently — no error is shown. For the content-injected keys (`orchestratorPromptDocs`, `workerPromptDocs`, `doctrineExtraPath`) a missing, unreadable, or empty file is skipped and nothing is injected. For `reviewPerspectivesPath` the pointer is omitted only when the file is missing — the file is not read at injection time, so an unreadable or empty file is still cited. Verify your paths after copying the example.
+
+## Trust
+
+Slate reads project configuration (`.pi/slate.json`) and injects project files (`orchestratorPromptDocs`, `workerPromptDocs`, `doctrineExtraPath`, `reviewPerspectivesPath`) **only in trusted projects**. In untrusted projects Slate runs with built-in defaults and injects nothing from the working tree.
+
+## Shipped docs
+
+In orchestrator mode, Slate appends a short **doctrine** (a block of numbered rules) to the orchestrator's system prompt each turn. The doctrine does not embed the workflow docs — it cites them by **absolute path**, resolved inside the installed package (not your project), and the orchestrator reads them on demand:
+
+- `docs/track-workflow.md` — the track-based workflow (research → design review → adversarial review → track review)
+- `docs/pr-publishing.md` — umbrella draft-PR publishing (cited only when `workflow.draftPRs` is `true`)
+- `docs/review-rules.md` — review discipline and finding rules
+- `docs/design-principles.md` — Slate's own design rationale
+
+Project-specific additions layer on top — they extend, not replace, the shipped doctrine — via two distinct mechanisms:
+
+- **Content injection**: `doctrineExtraPath` (appended to the doctrine itself, re-read at each prompt assembly) and `orchestratorPromptDocs` / `workerPromptDocs` (appended to the respective system prompts).
+- **Pointer**: `reviewPerspectivesPath` is cited by path from the doctrine's review rule and read on demand, like the shipped docs.
+
+## License
+
+Apache-2.0 — see [LICENSE](LICENSE).

package/docs/design-principles.md

@@ -0,0 +1,183 @@
+# Slate design principles
+
+Distilled from the Random Labs technical report introducing the Slate
+architecture ("thread weaving") — the open-beta release write-up of the
+Slate agent (published as the npm package `@randomlabs/slate`); no stable
+URL is recorded, so the distillation here is the in-repo source of truth
+for its content. This document records the reasoning behind the
+extension's design so maintainers — and the orchestrator itself, on
+demand — can check changes and behavior against the original intent.
+
+Note: module headers cite ExecPlan decision ids (D3–D9, M1–M3) from the
+original implementation plan, which is not part of this repository. This
+document is the in-repo source for the architecture rationale.
+
+## 1. The problems Slate is built to solve
+
+Three compounding problems in LLM agents — each tractable in isolation; the
+difficulty is that they interact:
+
+1. **Long-horizon tasks** — path-dependent tasks whose minimum successful
+   step count exceeds what a minimal tool-calling loop can do. Solving them
+   requires adequate working memory, a strategy/tactics balance, and the
+   ability to integrate information discovered mid-task without losing the
+   overall goal.
+2. **Working memory and the "Dumb Zone"** — models cannot attend uniformly
+   across the context window; retrieval quality degrades non-uniformly as
+   context grows ("context rot"). The usable prefix is working memory; the
+   degraded tail is the Dumb Zone. Context must be managed like scarce RAM.
+3. **Strategy vs tactics** — strategy is open-ended planning toward the
+   goal; tactics are learned local action sequences (run a command, extract
+   X from file Y). The AlphaGo/AlphaZero lineage architected this split
+   explicitly (value network = positional strategy, policy network = move
+   tactics), and probing shows tactical concepts are learned before
+   strategic ones. Software engineering spans the whole spectrum — a good
+   harness lets the model strategize without drowning in tactical detail.
+
+Supporting concepts:
+
+- **Knowledge overhang** — knowledge a model holds latently but cannot
+  access tactically without scaffolding (plans, chain-of-thought, harness
+  structure). Corollary: the bottleneck in long-horizon agentic work is
+  context management, not model intelligence.
+- **Expressivity** — a harness is expressive when few output operations can
+  reach many end states (a `sed`-only harness can read, write, and search;
+  a `file_read`-only harness can never edit). Rigid task graphs lower
+  expressivity.
+- **Inductive bias** — models default to interfaces they were trained on.
+  The harness builder's job is to make the desired behavior the natural
+  behavior.
+
+## 2. Why prior approaches fall short
+
+No prior approach solves all of the above simultaneously; each trades one
+or two problems for the others:
+
+- **Compaction** (sliding windows, Claude Code compaction, Amp handoffs) —
+  non-deterministically lossy: important information can vanish
+  unpredictably.
+- **Naive subagents** (Codex/Claude Code) — isolate context well, but
+  synchronize by message passing: the parent only gets a response string,
+  so information fails to cross the context boundary. Works mainly for
+  exploratory search over immutable data.
+- **Markdown plans** — force the model to strategize (tapping the knowledge
+  overhang) but go stale. Three failure modes: underspecified plans,
+  incomplete execution ("declaring victory early"), and forgetting to
+  update the plan when new information arrives.
+- **Direct task decomposition (task trees, gated steps)** — thorough and
+  resistant to early stopping, but rigid: adapting to new information means
+  rewriting the tree, and unintegrated subtask results get orphaned. Low
+  expressivity.
+- **RLM / recursive decomposition** — the right primitives (context by
+  reference, natural decomposition through a familiar interface), but
+  unbounded recursion needs a guard against over-decomposition, and REPL
+  execution yields no intermediate feedback: the model commits to N steps
+  blind and only learns the outcome at the end — no course correction in a
+  mutating environment.
+- **Strategize–delegate–compress stacks (Devin, Manus, Altera/PIANO)** —
+  every compress-and-return boundary risks dropping critical state, and the
+  strict planner/executor split adds inertia and reduces reactivity.
+- **ReAct** — maximally reactive and expressive but has no context
+  isolation, no compaction story, and no parallelism: the single context
+  fills until quality degrades.
+
+## 3. Slate's answer: threads, episodes, thread weaving
+
+- **Thread** — a persistent worker that executes ONE bounded action at a
+  time, then pauses and hands control back. Threads are general workers
+  serving the system's current intent, not purpose-specific personas. A
+  thread accumulates context across its actions, acting as a reusable store
+  for one work stream.
+- **Episode** — the compressed, structured record of the steps a thread
+  took to complete one action: important results retained, tactical trace
+  dropped. Episodes — not message passing — are the synchronization
+  primitive. Because a bounded action has a natural completion boundary,
+  compaction happens at meaningful moments instead of arbitrarily
+  mid-stream. This is a tractable form of episodic memory.
+- **Composability** — episodes are inputs: any thread can be initialized
+  with prior episodes (from any thread), inheriting conclusions without
+  inheriting full context. Context-by-reference routing is what
+  distinguishes threads from subagents that return a single string.
+- **Thread weaving** — the orchestrator dispatches, threads execute,
+  episodes compose. Decomposition is implicit and adaptive: the
+  orchestrator never commits to a static plan, but is forced to externalize
+  work as bounded, compressible units. Frequent bounded synchronization
+  gives per-episode feedback, so strategy updates mid-task instead of
+  failing at the end.
+- **OS framing** — the orchestrator is the kernel; threads are processes;
+  episodes are process return values committed into the kernel's working
+  memory; the context window is RAM — scarce and actively managed. Each
+  thread return is a scheduled opportunity to decide what is retained,
+  compressed, or discarded.
+
+## 4. Operating principles
+
+- **P1 — One dispatch, one bounded action.** An action is a tactic-sized
+  unit: clear, completable, verifiable.
+- **P2 — Episodes are the synchronization primitive.** No back-and-forth
+  message passing between orchestrator and workers.
+- **P3 — Compress at completion boundaries.** Compaction is built into the
+  action lifecycle, not applied as emergency lossy surgery.
+- **P4 — Compose context by reference.** Pass episode ids, not restated
+  content; the episode store stays the source of truth.
+- **P5 — Decompose implicitly and adaptively.** No upfront static plan;
+  update strategy after every episode; failed episodes demand adaptation,
+  not blind retry.
+- **P6 — Make desired behavior the natural behavior.** In orchestrator
+  mode, tactical tools are removed, so delegation is the only way to act.
+- **P7 — Guard against over-decomposition.** Workers cannot spawn threads
+  (depth-1); recursion is structurally impossible.
+- **P8 — Per-episode feedback beats blind N-step execution.** Bounded
+  actions return before the next decision, so course correction is always
+  possible.
+- **P9 — Parallelize independent actions.** Independent work streams run as
+  concurrent threads whose episodes are synthesized afterward.
+- **P10 — Treat context as RAM.** Budget it; when the budget is exceeded,
+  pause dispatching and hand off to a fresh session rather than degrading
+  in the Dumb Zone.
+
+## 5. Where each principle lives in the code
+
+| Principle | Implementation |
+|---|---|
+| P1 bounded actions | `tools.ts` `thread` tool contract; doctrine rule 1 in `mode.ts` |
+| P2 episodes as sync | `threads.ts` dispatch lifecycle returns an episode per action |
+| P3 boundary compression | `episodes.ts` episode compression on action completion |
+| P4 context by reference | `tools.ts` `context` parameter injects prior episodes by id |
+| P5 adaptive decomposition | doctrine rule 6 in `mode.ts`; no plan structure imposed anywhere |
+| P6 natural behavior | `mode.ts` `ORCHESTRATOR_TOOLS` restriction (read-only + slate tools) |
+| P7 over-decomposition guard | `worker.ts` recursion guard — workers load no extensions, so the `thread` tool never exists for them |
+| P8 per-episode feedback | `threads.ts` synchronous dispatch; episode returned to orchestrator |
+| P9 parallelism | `threads.ts` `maxConcurrent` queueing; doctrine rule 2 in `mode.ts` |
+| P10 context as RAM | `handoff.ts` context-budget auto-pause + fresh-session handoff |
+
+## 6. Runtime knowledge: what the orchestrator knows, and when
+
+The orchestrator's knowledge of these principles is two-tier, following
+the load-on-demand discipline the extension itself prescribes:
+
+- **Tier 1 — always loaded.** The doctrine in `mode.ts` is the operational
+  distillation of P1–P10, the review discipline defined in
+  `review-rules.md`, and a pointer to the track-based workflow protocol
+  that ships with this package (`track-workflow.md` in this directory),
+  appended to the system prompt every turn while orchestrator mode is
+  on. It costs a few hundred tokens and covers everything routine
+  dispatching needs.
+- **Tier 2 — on demand.** This document. The doctrine carries a short
+  pointer to it (doctrine rule 10); the orchestrator reads it only when
+  reasoning about the architecture itself — explaining slate, modifying
+  the extension, or making a non-obvious routing/compaction decision.
+
+The same discipline applies to consuming-project role guidance:
+`prompt-docs.ts` injects per-role guideline docs (orchestrator via
+`before_agent_start`, workers via `appendSystemPrompt`). Defaults are
+compiled into `prompt-docs.ts`; a consuming project supplies its own
+documents through the optional `slate.json` keys `orchestratorPromptDocs`
+/ `workerPromptDocs`. Each role's always-loaded surface carries only its
+own rules; the rest stays on demand.
+
+Injecting this full document every turn would be self-defeating: it would
+spend the working memory the architecture exists to protect, and most of
+its content (prior-approach analysis, background concepts) is rationale,
+not operational instruction. Keeping rationale on demand and rules always
+loaded is itself an application of P10.

package/docs/pr-publishing.md

@@ -0,0 +1,123 @@
+# Draft-PR publishing
+
+Umbrella draft PR mechanics for the track-based workflow (see
+track-workflow.md in this directory). This document applies ONLY when
+`workflow.draftPRs` is enabled in the project's `slate.json` (default:
+false). When it is disabled, the workflow creates no PR; the
+workflow-log lifecycle in that mode is owned by track-workflow.md
+§ Research log.
+
+## Creation
+
+The umbrella draft PR is created once the design has passed the user
+design review and the adversarial review, BEFORE implementation:
+
+- Created as a DRAFT, based on the repository's default development
+  branch.
+- If the working branch has no diff against the base yet, land a
+  bootstrap empty commit so the PR can be created.
+- At creation, the research log's content folds into the PR
+  description — Key decisions, Risks, and Open questions feed the
+  corresponding Planned-changes subsections, and the design review and
+  adversarial review verdict lines land in Risks & accepted
+  trade-offs — and the log file is deleted. Decisions made after PR
+  creation are appended to the PR description directly.
+- The user approves the description (and the track split — see
+  track-workflow.md) before implementation starts.
+
+## Description rules
+
+The description follows the repository's PR template, if any, and
+carries three parts: Motivation (why), "Planned changes" (detailed but
+high-level), and "Tracks" (a display table, multi-track changes only).
+
+Write Planned changes at a high design level using the MAIN DOMAIN
+ENTITIES from the code — real class and component names. Hard guards:
+no file paths, no method signatures. If a sentence would change when a
+method is renamed, it is too deep.
+
+Subsections activate when their content exists:
+
+- **Current state** — the before-picture per affected area.
+- **What changes** — the externally observable contract/behavior: API
+  surface, semantics, defaults, persisted formats, compatibility,
+  concurrency guarantees.
+- **How** — design-level description.
+- **Key decisions** — chosen vs rejected; preempts "why not X?" review
+  comments.
+- **Out of scope** — explicit non-goals.
+- **Risks & accepted trade-offs** — including the design review and
+  adversarial review verdict lines.
+- **Verification approach** — 1–2 lines.
+
+"Deep enough" test: a reviewer who knows the codebase but not this
+change can (1) predict which subsystems the diff touches, (2) evaluate
+each track's diff against a stated intent, and (3) answer "why not
+alternative X" without asking.
+
+The description becomes the squash-commit body on the default
+development branch — it is the permanent git-archaeology record for
+the change. Write it accordingly.
+
+## Tracks table
+
+The description's Tracks section holds the track table; its
+constraints (display-only, no SHAs, never the source of truth for
+track boundaries) are owned by track-workflow.md § Track table.
+Single-track changes carry an "N/A (single-track)" placeholder instead
+of a table.
+
+## Keeping the PR in sync
+
+Keep the title and description synchronized with what is actually
+pushed: update a track's table row when its marker commit lands,
+append post-design decisions as they are made, and revise Planned
+changes whenever reality diverges from it. A stale description fails
+the "deep enough" test.
+
+## Ready-for-review flip
+
+The umbrella PR is the ONLY PR this workflow ever merges (squash), and
+the merge is performed BY THE USER — the agent never merges it.
+Flipping the PR to ready-for-review is the agent's last act before
+handing the PR to the user; post-flip fix work and post-merge cleanup
+stay agent duties (below). The flip is gated by this checklist,
+executed in order:
+
+- Any layered peer-review process (see track-workflow.md § Peer
+  review) is completed or explicitly user-waived — flipping never
+  discards a pending review.
+- All commits landed since the last user-approved gate are presented
+  to the user.
+- Strip the whole Tracks section from the description, whatever its
+  form — the table for multi-track changes or the "N/A (single-track)"
+  placeholder — plus any notes under it. Track numbers are ephemeral
+  branch-life identifiers: after the squash-merge the marker commits
+  are gone, so track references would dangle in the default branch's
+  history. The rest of the description — Motivation, Planned changes —
+  stays: it becomes the squash-commit body.
+- Update the PR title and description to the final state of the
+  change: the title names what was actually delivered — preserving any
+  prefixes or markers the project's conventions require — and the
+  description, including the Planned changes section, describes the
+  change as implemented, folding in everything added, dropped, or
+  reshaped since the draft PR was opened.
+- Last, re-read the whole PR description end-to-end to confirm the
+  as-flipped text tells one consistent story.
+
+## After the flip
+
+The user may wait for CI green and/or peer-review completion and ask
+the agent to fix test failures or review observations. The agent lands
+fixes as normal commits, keeps the description in sync, and presents
+agent-landed commits to the user as they land. Commits pushed directly
+by reviewers are visible in the PR UI; the agent reconciles the
+description with them on its next task. The user's merge act is the
+final approval.
+
+## After the merge
+
+Any cleanup a layered peer-review process requires (closing its review
+PRs, deleting its pinned branches) is an agent duty, executed when the
+user reports the merge or a later session detects it, per that
+process's own rules.