@justjoshn/pi-pstack 0.0.0-stage → 0.15.36-pi.11

package/docs/parity-status.md

@@ -0,0 +1,113 @@
+# Parity status
+
+This page records what `pi-pstack` reproduces from Cursor `pstack` 0.15.36 and its dependencies, what it translates to a native Pi mechanism, and what remains a host difference. It is a status record, not a claim that every workflow has been observed on both hosts.
+
+## Sources
+
+- Cursor `pstack` and `cursor-team-kit`: `github.com/cursor/plugins` at `6ea74587e384e1a8035c09ae33a46fe3f881eb9b`. The plugin manifest reports pstack 0.15.36. `package.json.pstackSource.revision` pins `ae5650fe`, this port's capture base; `git log ae5650fe..6ea74587 -- pstack cursor-team-kit` is empty, so both commits name the same plugin content.
+- Installed Cursor CLI at the time of the reference capture: `2026.10.22-27dcf5f`.
+- Native implementation: this package, which pins its upstream provenance in `package.json.pstackSource`.
+
+A path and byte comparison of the tracked trees shows every pstack skill, both pstack agents, all 18 team-kit skills, both team-kit agents, and both team-kit rules in this package. Seventeen team-kit skill files are byte-identical; `workflow-from-chats` describes Pi paths instead of Cursor paths. The changed pstack files are host translations. One upstream paragraph has no counterpart: `poteto-help`'s Claude Code install path and its `docs/harness/claude-code.md` adapter, because this package targets Pi only. The two non-portable frontmatter keys Cursor ships (`icon`/`color` on `poteto-mode`, `paths` on `typescript-best-practices`) are kept under `metadata`, the Agent Skills extension point; the port reads both, and it attaches `typescript-best-practices` by path through `src/path-skills.ts`.
+
+## Native translations
+
+| Cursor dependency | Pi mechanism |
+| --- | --- |
+| Skill invocation and the slash menu | `pi.getCommands()` plus `/name` to `/skill:name` input rewriting, with TUI completions |
+| Custom Mode | `/mode <skill>`, a branch-persisted entry, the skill attached once on the entering request, and a path-naming reminder in a `pstack_mode` prompt section on every later request |
+| Plan mode | `/plan`, a reduced tool set, `bash_readonly` in an OS sandbox, a `create_plan` artifact under `plans/`, and `/plan build` |
+| Always-applied model rule | `pstack-models.md` read before every request and injected as `pstack_models` |
+| Task subagents | One `pi --mode rpc` process per agent, with resume, steer, fork, `fork_turns`, readonly, attachments, background notices, and local worktrees |
+| Agent types | The four bundled agent Markdown files, loaded into the child system prompt |
+| Project and agent stores | Workspace state folders named in the host system prompt |
+| Todo list | `TodoWrite` with branch reconstruction and a TUI widget |
+| Structured questions | `AskQuestion` over Pi select and input dialogs |
+| `/loop` | `Loop` tool and `/loop` command. A fixed interval is wall-clock from each fire, like Cursor's sleep loop, and accepts the interval first or last (`5m`, `10 minutes`, `every 10 minutes`). An interval-less loop is agent-paced through `Loop` `wake_in`. |
+| Path-scoped skills | `src/path-skills.ts` attaches a loaded skill once per session when the session reads or edits a file matching its `paths`/`globs` (also under `metadata`), the way Cursor files a skill under `fileGlobbed` |
+| Always-applied rules | The two shipped rule bodies are read into `pstack_host` verbatim on every request, not paraphrased |
+| Todo list state | `TodoWrite` writes branch entries, and the current list rides in the `pstack_todos` prompt section on every request, so a compacted context still sees it |
+| Automations | Durable timer, GitHub, and webhook subscriptions with a detached supervisor |
+| `/visualize` | `visualize` tool that returns a Unicode chart and writes an SVG |
+| Benny automation pack | The same text workflows, run through subscriptions instead of the Cursor Automations editor |
+
+## Reference observations
+
+A live Cursor CLI session with both plugins loaded produced these contracts.
+
+- A foreground `Task` call returned the child's final text, an agent ID, `isBackground: false`, and a duration.
+- A background `Task` call returned `isBackground: true` with no conversation steps, then delivered a completion message naming the title and agent ID. The parent turn resumed from that message.
+- `Task` arguments included `description`, `prompt`, `subagentType`, `model`, `agentId`, `attachments`, `mode`, `environment`, and `machine`.
+- The CLI's model-facing `Task` schema exposes only `description`, `prompt`, `model`, `resume`, `subagent_type`, `file_attachments`, and `run_in_background`. It does not expose `readonly`, `environment`, `cloud_base_branch`, `fork`, or `fork_turns`, so those pstack arguments are inert in that CLI. A live Task started with `readonly: true` wrote a file, which confirms the CLI did not enforce it.
+- Pi exposes those arguments and enforces `readonly`, because the pstack skills use it to keep explorers and reviewers from editing. That is a deliberate correction of the host gap, not literal reproduction of it.
+- A live Task asked for `environment: "cloud"` ran in the same working directory on the same machine, with `environment: SUBAGENT_EXECUTION_ENVIRONMENT_UNSPECIFIED`. In this CLI, the cloud argument was inert too. Pi's local worktree is a stronger isolation than that observation, though still weaker than the documented cloud VM.
+
+Pi returns the child's final text with a transcript path and usage, and reports background completion as a `pstack/task` notification. The payload shape differs because the hosts are different; the observable workflow, a subagent result arriving after the parent turn, is preserved.
+
+## Repairs in 0.15.36-pi.11
+
+Each code change has a regression in the package suite. Text restorations are recorded here and checked against the reference by the parity audit.
+
+- A background completion notice names the subagent by its `description`, the field the Task schema says names it in the UI, instead of 200 characters of prompt. Cursor's own completion block puts the child's prompt in its `title` field; this notice keeps both the name and a prompt excerpt.
+- A foreground Task streams its child's last tool steps and latest line into the running card, throttled and capped, instead of showing a blank wait until the child settles.
+- `/loop` accepts the interval at either end and in unit words (`10 minutes`, `every 10 minutes`), and a bare `/loop` prints usage. A fixed loop keeps wall-clock cadence, so a long run no longer pushes the next tick out.
+- A bare `/plan` enters plan mode, as Cursor's plan command does, and plan mode keeps `stop_agent` so a launched child can still be stopped.
+- Rules that declare `paths`/`globs` attach on Pi when a matching file is read or edited, once per session.
+- The always-applied engineering rules are injected as their shipped text, including "interface fields" and "strict circular-dependency reason".
+- The todo list is in every request's system prompt, so compaction and a branch switch do not leave the model without it.
+- `cloud_base_branch` starts from `origin/<branch>` when that ref exists, as a cloud worker sees only pushed code, and falls back to a local ref.
+- A resume that names a `subagent_type` keeps the recorded type and says the parameter was ignored.
+- A detached `new_session` fire records its own final message as the run output, so `list_subscriptions` shows what the run said instead of `detached run completed`.
+- benny restores the upstream requirement that the pack and referenced configuration are committed before a subscription is enabled, and a sweep reads from the last successful sweep so downtime does not lose reports.
+- The architect runner prompt restores "What matters is independence between candidates."
+- `/ketchup` describes the `visualize` tool as a bar or line chart and delivers screenshots and video the way Pi can (read the image, open the video, name the path).
+- A webhook routine accepts `sessionStrategy: "new_session"`, so an accepted POST starts a fresh detached run with no chat open, the way a Cursor Automation runs, and records its outcome in `list_subscriptions`. Records written before this field existed still load and behave as `wake_self`. Evidence: `test/e2e/subscription-owner.e2e.test.ts` runs the real Pi CLI with the owner closed, POSTs to the bound listener, and finds the detached run's file and its recorded outcome.
+- A timer or webhook routine no longer expires after seven days. Cursor's automation docs define no expiry and describe "always-on agents"; the closest documented cap, 180 days, belongs to cloud agent subscriptions, a different surface. `expiresInHours` still sets one, `0` means none, and a GitHub watch keeps its 7-day default because it is this port's own surface.
+- Each workspace binds its own routines port. With no explicit port, a busy default falls back to an ephemeral port that the workspace reuses and reports in the routine URL, so two workspaces serve routines at once. An explicit `PSTACK_ROUTINES_PORT` still reports a collision as an error. Evidence: the two-workspace case runs two real Pi sessions and supervisors in `test/e2e/subscription-owner.e2e.test.ts`.
+- A mode now matches Cursor's injection. The skill body rides the first request after entry as the same `<skill>` block Pi's own `/skill:name` expansion builds, addressed with Cursor's entry notice, and every later request carries Cursor's reminder shape, which names the skill file. The mode re-sends the body after a restart. Evidence: `test/e2e/mode.e2e.test.ts` reads the real system prompt on the entering turn, on the next turn, and after a restart.
+- Plan mode saves an artifact. A `create_plan` tool writes the plan to `<pstack dir>/plans/<title>-<id8>.plan.md`, the file name shape Cursor uses, and records it on the branch. `/plan build` leaves plan mode and sends the file's content as the next turn, which is Cursor's "Yes, build locally" path without its review menu. Evidence: `test/e2e/plan.e2e.test.ts` drives the real CLI through `/plan`, the tool call, and `/plan build`.
+
+## Repairs in 0.15.36-pi.10
+
+Each code change has a regression in the package suite. Text restorations are recorded here and checked against the reference by the parity audit.
+
+- The root chat can resume and fork a descendant that a worktree worker started inside that worktree. Unrelated directories remain refused. New records store the owning workspace root; older records are recognized by this workspace's worktree folder.
+- A steering message stranded at settlement is resumed with its recorded images and keeps its own message boundary, instead of joining cleared text and dropping attachments.
+- A full subscription queue refuses a new event with HTTP 503 and `Retry-After` instead of evicting an event already acknowledged with 200.
+- Accepted `new_session` work stays eligible after expiry, so expiry stops new occurrences without stranding accepted work.
+- Task's settle-time notification appends to entries another extension proposed at the boundary instead of replacing them.
+- A `new_session` timer reuses the model, thinking level, and project trust it was created with, whether its owner is live or detached.
+- The upstream `ci-watcher` `model: fast` is restored; `fast` picks the parent model's listed `-fast` variant, else the parent model.
+- A pre-epoch supervisor lock is signaled only when its command line identifies this supervisor, so a reused pid cannot terminate an unrelated process.
+
+## Remaining host differences
+
+These are not closed by extension code.
+
+- **Background workers stop when the chat exits.** A Pi subagent is a child process of the chat, so quitting or shutting the session stops the agents it started and reports how many. Whether a Cursor background subagent outlives its parent is unobserved; the CLI changelog says cancelling the prompt stops its subagents, and only the cloud-agent product promises work after the laptop closes. A `new_session` subscription fire is Pi's one detached path, and it starts a fresh run rather than continuing a subagent.
+- **A mode carries its body once, then a reminder.** Cursor attaches the skill body on the turn that enters the mode and sends a path-naming reminder every turn after, in a user-message field. Pi rebuilds its prompt per request, so the body rides the first request after entry as a `pstack_mode` section and later requests carry the reminder. A process restart re-sends the body once, since the new process cannot know the conversation already saw it.
+- **A fork needs an idle source.** Cursor exposes `fork` only on its internal and cloud-agent protos, never on the CLI Task schema, and no doc states whether a fork may start while the source runs. A Pi fork runs in the source's directory or worktree, so a running source is refused; the orchestrate playbook says to wait for its completion notification.
+- **Plan mode's restriction is stricter than the host's.** Cursor's inspected CLI carries no client-side plan-mode filter, and its docs recommend stating read-only intent in the prompt, so the restriction on that surface is prompt or backend side; no source shows a plan-mode network rule. Pi enforces a tool allowlist and a sandboxed shell that also denies network, so `gh`, `curl`, and package queries can fail while planning.
+- **No host-managed plan review UI.** The plan is a file the agent writes with `create_plan`, and `/plan build` starts the work from it. Cursor also renders a plan review menu with "build locally", "build in cloud", and "no, propose changes", and can hand the plan to a cloud agent. Pi has no equivalent UI, and its plan review is the file plus the reply.
+- **benny's default cadence still needs its owning session open.** `wake_self` timers and webhook routines park while the owner is offline. A timer or webhook routine created with `sessionStrategy: "new_session"` runs detached with no chat open, but a detached run is one-shot: it must do the work in that run, because a background `Task` ends with it. Cursor automations run in the cloud and can hand work to anything. So benny keeps `wake_self` by default and setup names the tradeoff before switching.
+- **A routines listener is local, one port per workspace.** Cursor hosts each routine URL. Pi binds a loopback port per workspace, so two workspaces serve routines at once only because the default port falls back to an ephemeral one that the workspace then reuses and reports. A sender outside the machine still needs its own forwarding path.
+- **Cloud execution is local.** Cursor documents cloud VMs for its cloud product, but the inspected CLI schema omits `environment` and a requested cloud Task ran on the same machine and directory. Pi's `environment: "cloud"` goes further with a git worktree and a separate local process. Neither runs with the laptop off or provides a remote VM.
+- **Webhook routines are local.** Cursor provides a hosted routine endpoint and an editor panel. Pi serves an authenticated listener on loopback and stores a sender key in a local file. A sender outside the machine needs its own forwarding path.
+- **Slack-triggered automations need an external sender.** The Benny workflows can sweep on a timer or receive a local webhook. The Cursor Automations product and its Slack trigger are not reproduced.
+- **Inline video is not rendered in the reply.** The ketchup-video workflow writes the video and names its path; Pi opens it in the system player instead.
+- **Video attachments have no Pi model input.** Cursor can attach video to a Task. Pi names the video path for the agent instead of decoding the file as text; the agent opens it with a media tool.
+- **Hosted guidance and pages are outside the package.** MCP providers, ticket trackers, browsers, simulators, and Tailscale remain the user's own dependencies, as they are for Cursor.
+- **Background usage cannot enter parent totals.** Pi exposes no API that adds a later notification's usage to the parent session. It stays in the notification details and the child transcript.
+
+## Open evidence
+
+The following were not observed end to end and are not claimed.
+
+- A real model following every ported skill on both hosts.
+- Authenticated MCP exchanges and provider-specific tool behavior.
+- A successful live Cursor run for each workflow after the reference capture.
+- Reboot survival through a real launchd or systemd unit on both platforms.
+- Multi-workspace webhook routing.
+- Exactly-once external actions in Benny workflows.
+
+Passing package tests show the extension contracts hold. They do not replace those observations.

package/licenses/cursor-team-kit.LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 Cursor
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

package/licenses/pstack.LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 Lauren Tan
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

package/package.json

@@ -1,6 +1,88 @@
 {
   "name": "@justjoshn/pi-pstack",
-  "version": "0.0.0-stage",
-  "stub": true,
-  "description": "Temporary package placeholder for staged publishing"
-}
+  "version": "0.15.36-pi.11",
+  "description": "pstack 0.15.36 for Pi with its soft-dependency team-kit skills vendored in: rigorous agent workflows and principles as Pi skills, with the host features they use: subagents, todos, modes, plan mode, the model rule, /loop, subscriptions and routines, AskQuestion, and visualize",
+  "type": "module",
+  "license": "MIT",
+  "pstackSource": {
+    "repository": "https://github.com/cursor/plugins",
+    "revision": "ae5650febf4320793844c12c4adaa937f417f8da",
+    "path": "pstack",
+    "version": "0.15.36",
+    "lastChange": "c742dd61706fbacc1b95f1a20b7b988cf096ce6c"
+  },
+  "repository": {
+    "type": "git",
+    "url": "git+https://github.com/just-joshn/pi-customizations.git",
+    "directory": "extensions/pi-pstack"
+  },
+  "homepage": "https://github.com/just-joshn/pi-customizations/tree/main/extensions/pi-pstack#readme",
+  "bugs": {
+    "url": "https://github.com/just-joshn/pi-customizations/issues"
+  },
+  "keywords": [
+    "pi-package",
+    "pstack"
+  ],
+  "bin": {
+    "pstack-routines-supervisor": "./bin/pstack-routines-supervisor"
+  },
+  "engines": {
+    "node": ">=22.19.0"
+  },
+  "publishConfig": {
+    "access": "public",
+    "tag": "latest"
+  },
+  "pi": {
+    "extensions": [
+      "./src/index.ts"
+    ],
+    "skills": [
+      "./skills"
+    ]
+  },
+  "files": [
+    "src",
+    "skills",
+    "agents",
+    "automations",
+    "bin",
+    "docs",
+    "assets",
+    "licenses",
+    "packaging",
+    "README.md",
+    "CHANGELOG.md",
+    "LICENSE"
+  ],
+  "scripts": {
+    "typecheck": "tsc --noEmit -p tsconfig.json",
+    "test": "vitest run",
+    "test:coverage": "vitest run --coverage",
+    "test:statechart": "node --disable-warning=ExperimentalWarning --test skills/statechart/scripts/test/*.test.ts",
+    "live:e2e": "node scripts/live-e2e.mjs"
+  },
+  "dependencies": {
+    "@anthropic-ai/sandbox-runtime": "0.0.79"
+  },
+  "peerDependencies": {
+    "@earendil-works/pi-ai": "*",
+    "@earendil-works/pi-agent-core": "*",
+    "@earendil-works/pi-coding-agent": "*",
+    "@earendil-works/pi-tui": "*",
+    "typebox": "*"
+  },
+  "devDependencies": {
+    "@earendil-works/pi-ai": "1.5.0",
+    "@earendil-works/pi-agent-core": "1.5.0",
+    "@earendil-works/pi-coding-agent": "1.5.0",
+    "@earendil-works/pi-tui": "1.5.0",
+    "@types/node": "^22.19.0",
+    "@vitest/coverage-v8": "^5.0.2",
+    "fast-check": "4.5.3",
+    "typebox": "1.3.27",
+    "typescript": "^7.0.2",
+    "vitest": "^5.0.2"
+  }
+}

package/packaging/install-supervisor-unit.sh

@@ -0,0 +1,164 @@
+#!/usr/bin/env bash
+# Install a per-workspace user unit so pstack-routines-supervisor survives reboot.
+# Idempotent: re-run overwrites the same unit path and reloads it.
+set -euo pipefail
+
+usage() {
+  echo "Usage: $0 --workspace <abs-path> [--bin <abs-path>]" >&2
+  exit 2
+}
+
+workspace=""
+bin=""
+while [[ $# -gt 0 ]]; do
+  case "$1" in
+    --workspace)
+      [[ $# -ge 2 ]] || usage
+      workspace=$2
+      shift 2
+      ;;
+    --bin)
+      [[ $# -ge 2 ]] || usage
+      bin=$2
+      shift 2
+      ;;
+    -h|--help)
+      usage
+      ;;
+    *)
+      usage
+      ;;
+  esac
+done
+
+[[ -n "$workspace" ]] || usage
+[[ "$workspace" = /* ]] || {
+  echo "error: --workspace must be an absolute path" >&2
+  exit 2
+}
+[[ -d "$workspace" ]] || {
+  echo "error: workspace is not a directory: $workspace" >&2
+  exit 1
+}
+
+here=$(cd "$(dirname "$0")" && pwd)
+if [[ -z "$bin" ]]; then
+  bin=$(cd "$here/../bin" && pwd)/pstack-routines-supervisor
+fi
+[[ "$bin" = /* ]] || {
+  echo "error: --bin must be an absolute path" >&2
+  exit 2
+}
+[[ -x "$bin" ]] || {
+  echo "error: supervisor bin missing or not executable: $bin" >&2
+  exit 1
+}
+
+# Short content hash so labels stay stable across re-runs for the same path.
+safe_hash=$(printf '%s' "$workspace" | shasum -a 256 | cut -c1-12)
+os=$(uname -s)
+
+install_darwin() {
+  local label="com.pstack.routines-supervisor.${safe_hash}"
+  local plist_dir="$HOME/Library/LaunchAgents"
+  local plist="$plist_dir/${label}.plist"
+  local uid domain
+  uid=$(id -u)
+  domain="gui/${uid}"
+
+  mkdir -p "$plist_dir"
+  cat >"$plist" <<EOF
+<?xml version="1.0" encoding="UTF-8"?>
+<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
+<plist version="1.0">
+<dict>
+  <key>Label</key>
+  <string>${label}</string>
+  <key>ProgramArguments</key>
+  <array>
+    <string>${bin}</string>
+    <string>--workspace</string>
+    <string>${workspace}</string>
+  </array>
+  <key>WorkingDirectory</key>
+  <string>${workspace}</string>
+  <key>KeepAlive</key>
+  <true/>
+  <key>RunAtLoad</key>
+  <true/>
+  <key>StandardOutPath</key>
+  <string>/tmp/pstack-routines-supervisor-${safe_hash}.log</string>
+  <key>StandardErrorPath</key>
+  <string>/tmp/pstack-routines-supervisor-${safe_hash}.err</string>
+</dict>
+</plist>
+EOF
+
+  launchctl bootout "${domain}/${label}" 2>/dev/null || true
+  if launchctl bootstrap "$domain" "$plist" 2>/dev/null; then
+    launchctl enable "${domain}/${label}" 2>/dev/null || true
+    launchctl kickstart -k "${domain}/${label}" 2>/dev/null || true
+  else
+    # Older launchctl: load -w is enough for KeepAlive + RunAtLoad.
+    launchctl unload -w "$plist" 2>/dev/null || true
+    launchctl load -w "$plist"
+  fi
+
+  echo "Installed launchd agent: $plist"
+  echo "  label: $label"
+  echo "  bin: $bin"
+  echo "  workspace: $workspace"
+  echo "Uninstall:"
+  echo "  launchctl bootout ${domain}/${label} 2>/dev/null || launchctl unload -w \"$plist\""
+  echo "  rm -f \"$plist\""
+}
+
+install_linux() {
+  local unit_name="pstack-routines-supervisor-${safe_hash}.service"
+  local unit_dir="$HOME/.config/systemd/user"
+  local unit="$unit_dir/$unit_name"
+
+  mkdir -p "$unit_dir"
+  cat >"$unit" <<EOF
+[Unit]
+Description=pstack durable routines supervisor (${safe_hash})
+After=default.target
+
+[Service]
+Type=simple
+WorkingDirectory=${workspace}
+ExecStart=${bin} --workspace ${workspace}
+Restart=on-failure
+RestartSec=3
+
+[Install]
+WantedBy=default.target
+EOF
+
+  systemctl --user daemon-reload
+  systemctl --user enable --now "$unit_name"
+
+  echo "Installed systemd user unit: $unit"
+  echo "  unit: $unit_name"
+  echo "  bin: $bin"
+  echo "  workspace: $workspace"
+  echo "If no login session is open after reboot, enable linger so the user unit still starts:"
+  echo "  loginctl enable-linger \"$(id -un)\""
+  echo "Uninstall:"
+  echo "  systemctl --user disable --now $unit_name"
+  echo "  rm -f \"$unit\""
+  echo "  systemctl --user daemon-reload"
+}
+
+case "$os" in
+  Darwin)
+    install_darwin
+    ;;
+  Linux)
+    install_linux
+    ;;
+  *)
+    echo "error: unsupported OS: $os (need Darwin or Linux)" >&2
+    exit 1
+    ;;
+esac

package/packaging/launchd/com.pstack.routines-supervisor.plist.example

@@ -0,0 +1,24 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
+<plist version="1.0">
+<dict>
+  <key>Label</key>
+  <string>com.pstack.routines-supervisor</string>
+  <key>ProgramArguments</key>
+  <array>
+    <string>/absolute/path/to/node_modules/@justjoshn/pi-pstack/bin/pstack-routines-supervisor</string>
+    <string>--workspace</string>
+    <string>/absolute/path/to/your/repo</string>
+  </array>
+  <key>WorkingDirectory</key>
+  <string>/absolute/path/to/your/repo</string>
+  <key>KeepAlive</key>
+  <true/>
+  <key>RunAtLoad</key>
+  <true/>
+  <key>StandardOutPath</key>
+  <string>/tmp/pstack-routines-supervisor.log</string>
+  <key>StandardErrorPath</key>
+  <string>/tmp/pstack-routines-supervisor.err</string>
+</dict>
+</plist>

package/packaging/systemd/pstack-routines-supervisor.service.example

@@ -0,0 +1,15 @@
+# Does not require root. Reboot survival needs linger enabled for the user when no login session is open.
+
+[Unit]
+Description=pstack durable routines supervisor
+After=default.target
+
+[Service]
+Type=simple
+WorkingDirectory=/absolute/path/to/your/repo
+ExecStart=/absolute/path/to/node_modules/@justjoshn/pi-pstack/bin/pstack-routines-supervisor --workspace /absolute/path/to/your/repo
+Restart=on-failure
+RestartSec=3
+
+[Install]
+WantedBy=default.target

package/skills/architect/SKILL.md

@@ -0,0 +1,79 @@
+---
+name: architect
+description: "Sketch types, signatures, and module structure before code, then stay in the loop while implementation fills in. Use for /architect, 'architect this', 'design this', or non-trivial work where jumping to code would lock in the wrong shape."
+disable-model-invocation: true
+---
+
+# Architect
+
+Sketch types, function signatures, class shapes, and module boundaries with `not implemented` bodies and pseudocode. Synthesize across multiple model perspectives, then fill in code against the chosen sketch.
+
+## Start
+
+Open a todolist with one entry per phase before starting.
+
+1. Ground
+2. Sketch
+3. Agree
+4. Implement
+5. Scrap
+
+## Phase A: Ground the problem
+
+Build a real mental model of every system the new code touches. Run the **how** skill over the relevant subsystems. Naming a file isn't grounding. Produce the architectural explanation `how` prescribes. If the design redefines ownership or layering, also run the **why** skill on the existing shape so the rationale becomes a constraint, not a guess.
+
+Skip Phase A only when the work is genuinely greenfield with no surrounding system to integrate.
+
+## Phase B: Sketch
+
+Run the **arena** skill with the design-sketch task and the Phase A grounding artifacts. Pass `references/runner-prompt.md` as each runner's prompt.
+
+Take the runners from the `architect runners` line in the model rule at the path the system prompt names, in place of the `arena runners` line. If the rule or that line is missing, use `claude-opus-5-5:xhigh` and `grok-4.7[effort=xhigh,fast=true]`. Alias and rejected entries follow the runner rules in the **arena** skill's Phase A.
+
+Design it twice. Require at least two structurally distinct candidates before synthesis, even when the first looks sufficient. This is the **exhaust-the-design-space** principle skill made concrete. Whole-shape alternatives, not point fixes inside one shape.
+
+Screen every candidate against [`references/design-red-flags.md`](references/design-red-flags.md) before synthesis. Assume the next contributor is an agent that sees only the files it opened, copies the nearest example, and takes the shortest path that compiles. Prefer the design where a change that looks right from one file is right for the whole repo.
+
+Compare viable candidates on interface depth. Prefer the design that hides more complexity behind a smaller, simpler public surface. A rich interface can keep call chains short by concentrating capability instead of scattering it across layers.
+
+Arena returns one synthesized design package. Its synthesis decision fills the rationale's "Synthesis decision" section.
+
+## Phase C: Agree (opt-in)
+
+Default: proceed directly to implementation with the synthesized design. No human checkpoint.
+
+Opt in to a checkpoint when the invoker explicitly asks: "/architect with checkpoint," "stop and show me before implementing," or similar. Then surface the synthesized design and pause for sign-off.
+
+The synthesis can ship as its own commit either way, as the "scaffold first" mode of the **foundational-thinking** principle skill. Planned and scoped breakage during fill-in is fine, per the **outcome-oriented-execution** principle skill. For adversarial pressure on the design before implementing, run the **interrogate** skill on the synthesized sketch.
+
+If the human pushes back on the shape (in a checkpoint or after the fact), treat that as Phase A evidence. Re-ground and re-run Phase B before writing more code.
+
+## Phase D: Implement against the sketch
+
+Replace `not implemented` bodies with code, pseudocode with logic. The synthesized sketch is the contract. Deviations from it are signal to surface, not friction to absorb silently. If a function needs a parameter the sketch didn't anticipate, ask whether the sketch was wrong, the requirement was missed, or the implementation is overreaching.
+
+## Phase E: Scrap when the architecture is wrong
+
+If implementation keeps producing friction the sketch can't absorb, throw the sketch out. Don't bolt fixes onto a wrong design, per the **redesign-from-first-principles** and **fix-root-causes** principle skills.
+
+The signal is a *pattern*, not single instances. Tells:
+
+- The same shape of workaround appearing repeatedly across unrelated code.
+- Multiple unrelated edge cases that all need special-case branches.
+- Types that need escape hatches (`any`, casts, optional fields always set in practice) to compile.
+- The "we need a lock" reflex when the sketch said the state wasn't shared.
+- Callers having to know the abstraction's internal rules to use it.
+- Two or more independent Phase D deviations of the same shape across the implementation.
+
+Use judgment. A few edge cases don't condemn an architecture. Some problems are legitimately complex. Complexity in the data is not complexity in the design.
+
+When you scrap:
+
+1. Re-run the **how** skill over what's been built.
+2. Redesign as if the new constraints had been day-one assumptions, per redesign-from-first-principles.
+3. Subtract before adding, per the **subtract-before-you-add** principle skill. The new sketch should be smaller than the old one before it grows.
+4. Return to Phase B and re-run arena.
+
+## Outputs
+
+The caller's usage is written first and the type sketch derived from it. One file with new types and signatures for small changes. Module map plus type definitions for larger work. The rationale ships alongside, shaped per `references/rationale-template.md`, including the usage sketch and the synthesis decision.

package/skills/architect/references/design-red-flags.md

@@ -0,0 +1,57 @@
+# Design red flags
+
+Screen every candidate before synthesis. A red flag is a reason to revise or reject the shape.
+
+## Shallow module
+
+A shallow module exposes a large interface while hiding little complexity. Judge depth by the capability and policy hidden behind the public surface relative to the size of that surface. Prefer a simple interface backed by substantial behavior.
+
+Do not confuse a deep module with a deep call chain. A deep call chain scatters understanding across layers. A deep module concentrates capability behind one interface.
+
+Look for these signs:
+
+- Callers coordinate several methods to complete one operation.
+- Public options expose internal stages or implementation choices.
+- Learning the interface does not save the caller from learning the implementation.
+
+## Information leakage
+
+Information leakage makes multiple modules depend on the same internal decision. A representation, policy, or protocol detail appears in more than one place, so changing it requires coordinated edits.
+
+Public re-exports of transport or wire types are leakage. Parse external data into domain types behind the interface. Keep storage schemas, framework objects, and protocol details private.
+
+## Temporal decomposition
+
+Temporal decomposition organizes modules by execution order instead of the knowledge they own. Separate load, validate, transform, and save stages often repeat one representation and its invariants across several boundaries.
+
+Group code around domain knowledge and ownership. Methods that run at different times can still belong to one module when they protect the same decisions.
+
+## Pass-through method
+
+A pass-through method forwards the same arguments to another method with the same shape. It adds a layer without hiding complexity.
+
+Remove it or move responsibility to the module that can complete the operation. Keep a forwarding boundary only when it adds policy, adaptation, or a distinct abstraction.
+
+## Split ownership
+
+More than one module writes the same state or keeps its own copy of it. An agent that edits one writer can't see the others, so their rules diverge.
+
+Give each piece of state one owner. Other modules read it or ask the owner to change it.
+
+## Two ways to do one task
+
+The design supports more than one way to do the same task. An agent copies whichever way it finds first, so every way keeps gaining callers.
+
+Keep one way. Move callers off the others and delete them in the same change.
+
+## Importable internals
+
+A caller can import a module's internals. An agent takes the shortest path that compiles, so it imports them directly and they become part of the interface.
+
+Make internals unreachable from outside the module, so an import from outside fails the build.
+
+## Hand-synced list
+
+Two or more places list the same items, and adding an item means editing every list. An agent that sees one list updates only that one.
+
+Keep one list and derive the others from it. If a list can't be derived, make the build fail when the lists disagree.

package/skills/architect/references/rationale-template.md

@@ -0,0 +1,35 @@
+# Rationale template
+
+The prose that ships alongside the type sketch. One page. Sentence-case headings, no boilerplate. Replace the italic notes with actual content.
+
+## Problem
+
+*One paragraph. What we're trying to do, and what about the existing system or constraints makes the shape non-obvious. If [Phase A](../SKILL.md#phase-a-ground-the-problem) surfaced constraints the design must honor (existing types to interop with, callers we can't break, invariants that crossed our boundary), name them here so the reader sees the same constraints you saw.*
+
+## Usage (caller's view)
+
+*Write this first, before the type sketch. Show the README or quickstart the consumer reads, plus two or three realistic call sites in their own code. What they import, what they call, what comes back. The type sketch in [Shape](#shape) is derived from this. The two must agree. When they diverge, reconcile the sketch to the usage, not the reverse. The caller's experience is the spec. The types serve it.*
+
+## Shape
+
+*The recommended architecture. Data structures first. Then how data flows through the signatures. Name the load-bearing decisions. State which invariants are encoded in types, where validation lives, and what the system deliberately does not do. Judge interface depth explicitly. State what complexity the public surface hides, what remains exposed to callers, and why the interface is no larger than needed. Cite the principle behind each decision (e.g., `per boundary-discipline`). Don't restate it.*
+
+## Synthesis decision
+
+*Filled in by [arena](../../arena/SKILL.md). Records which candidate became the base and why, what was adapted from each of the others, and what was rejected and why.*
+
+## Tradeoffs accepted
+
+*One bullet per tradeoff the chosen shape makes. Form: "we accept X in exchange for Y." Name anything a future reader might mistake for an oversight, including things that look like premature optimization or premature simplification.*
+
+## Alternatives considered
+
+*Required. Name at least one concrete alternative shape, with one line on why it lost. Judge each alternative on interface depth, not implementation simplicity alone. Name the complexity it exposes to callers and the complexity it hides. Two or three alternatives belong here when the design space had real contenders. One is fine when the constraints forced the answer, with the conclusion phrased as "this was the only viable shape because..." Avoid listing flavors of the same shape. This section covers design alternatives the chosen shape considered and rejected, not other runner candidates.*
+
+## Open questions and risks
+
+*Things you noticed during the sketch that the human needs to weigh in on, and risks worth flagging before implementation starts. Phrase as questions, not assertions, so the human's answer is the resolution rather than a comment.*
+
+## Next implementation step
+
+*The first thing to build against the sketch. One sentence. What you'd start writing immediately after synthesis (or after Phase D sign-off, if a checkpoint was opted into).*