@justjoshn/pi-pstack 0.15.36-pi.1 → 0.15.36-pi.7

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

@@ -28,7 +28,7 @@ Skip Phase A only when the work is genuinely greenfield with no surrounding syst
 
 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 (`~/.pi/agent/pstack/pstack-models.md`), 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:xhigh`. Alias and rejected entries follow the runner rules in the **arena** skill's Phase A.
+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:xhigh`. 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.
 

package/skills/architect/references/runner-prompt.md

@@ -1,6 +1,6 @@
 # Architect runner prompt
 
-The orchestrator passes this file through to every parallel candidate runner during Phase B and fills in the variable inputs around it: the task, the Phase A grounding artifacts, the isolated working directory, and the path to write outputs. The working directory is a git worktree when available, otherwise a per-runner subdirectory under the sketch dir. What matters is independence between candidates.
+The orchestrator passes this file through to every parallel candidate runner during Phase B and fills in the variable inputs around it: the task, the Phase A grounding artifacts, the isolated working directory, and the path to write outputs. The working directory is a git worktree when available, otherwise a per-runner subdirectory under the sketch dir. Worktrees and scratch directories share the host machine. Use the distinct ports, output directories, and writable cache paths assigned in the brief for concurrent runtime candidates. Report a missing assignment before starting a conflicting process. If the design requires stronger machine isolation, report it as unavailable.
 
 You are producing one candidate design in architect's parallel exploration. Read the **architect** skill in full first. That's the workflow you're inside. Output a candidate design package: type sketch, function signatures, module map, and prose rationale shaped per [`rationale-template.md`](rationale-template.md).
 

package/skills/arena/SKILL.md

@@ -14,8 +14,8 @@ Every candidate gets the same prompt, so the prompt is the contract.
 
 1. State the artifact each candidate is producing.
 2. Derive the rubric, 3-6 concrete gradeable criteria for what success looks like on *this* task. It is the picker's tool in Phase D. Candidates only see the task.
-3. Pick the runners. Use the `arena runners` line in `~/.pi/agent/pstack/pstack-models.md`. If the rule or that line is missing, default to one each on `claude-opus-5-5:xhigh` and `grok-4.7:xhigh`. An `auto` or `inherit-parent` entry in this line means the parent model, so omit `model` for it. If the Task tool rejects a configured entry, run that seat on its family's default and say so. Families go by the model id after the provider: `claude-*` and `grok-*`. If the family has no default, or the Task tool rejects that default, use the closest valid slug of the same family from its error message, or drop the seat and say so when there is none. If dropped seats leave fewer than two runners, add runners until there are two, or one when the line has one entry. Run each added runner on a remaining runner's model, or omit `model` so it runs on the parent model when none remain, and say so. Spawn more runners when the arena covers multiple design directions. Use the same model N times when the work is generation-bound rather than judgment-sensitive. Pick each subagent's reasoning level (the `:<level>` suffix), and fast or standard speed when the model has a `-fast` variant, for the job, per the Subagents section of the **poteto-mode** skill.
-4. Give each candidate its own output location (a git worktree where possible, otherwise `/tmp/arena-<slug>/candidate-<n>/`), per the **separate-before-serializing-shared-state** principle skill.
+3. Pick the runners. Use the `arena runners` line in the model rule at the path the system prompt names. If the rule or that line is missing, default to one each on `claude-opus-5-5:xhigh` and `grok-4.7:xhigh`. An `auto` or `inherit-parent` entry in this line means the parent model, so omit `model` for it. If the Task tool rejects a configured entry, run that seat on its family's default and say so. Families go by the model id after the provider: `claude-*` and `grok-*`. If the family has no default, or the Task tool rejects that default, use the closest valid slug of the same family from its error message, or drop the seat and say so when there is none. If dropped seats leave fewer than two runners, add runners until there are two, or one when the line has one entry. Run each added runner on a remaining runner's model, or omit `model` so it runs on the parent model when none remain, and say so. Spawn more runners when the arena covers multiple design directions. Use the same model N times when the work is generation-bound rather than judgment-sensitive. Pick each subagent's reasoning level (the `:<level>` suffix), and fast or standard speed when the model has a `-fast` variant, for the job, per the Subagents section of the **poteto-mode** skill.
+4. Give each candidate its own output location (a git worktree where possible, otherwise `/tmp/arena-<slug>/candidate-<n>/`), per the **separate-before-serializing-shared-state** principle skill. Worktrees share this machine. Assign distinct ports, output directories, and writable cache paths for concurrent runtime candidates. If stronger machine isolation is required, report it as unavailable.
 
 ## Phase B: Fan out
 
@@ -25,7 +25,7 @@ If a candidate produces no output, proceed with N-1 and note the dropout.
 
 ## Phase C: Cross-judge
 
-After all Phase B candidates complete, choose one model from the `arena cross-judge pool` line in `~/.pi/agent/pstack/pstack-models.md`. If the rule or that line is missing, choose from `claude-opus-5-5:xhigh` and `grok-4.7:xhigh`. Spawn one readonly judge subagent. The judge is never from the parent's family, as Phase A step 3 defines families. Try only the entries from another family, in pool order. An `auto` or `inherit-parent` entry is the parent model, so it never judges. With the default pool, a Claude parent gets `grok-4.7:xhigh`. A parent from any other family gets `claude-opus-5-5:xhigh` first. If the Task tool rejects an entry, use a valid slug of the same model at the closest level or speed from the error message, or move to the next entry when there is none. If no entry from another family runs, try the default pool's entries from another family the same way. If none of those runs, skip the judge, pick the base alone in Phase D, and say so. If you do not know the parent's model, try the pool's real slugs in pool order and say that the family check did not run. The judge sees the rubric and the candidates by path label, scores each criterion, and recommends a base with rationale. It runs in parallel with your reading in Phase D. Don't spawn it while candidates are still writing.
+After all Phase B candidates complete, choose one model from the `arena cross-judge pool` line in the model rule at the path the system prompt names. If the rule or that line is missing, choose from `claude-opus-5-5:xhigh` and `grok-4.7:xhigh`. Spawn one readonly judge subagent. The judge is never from the parent's family, as Phase A step 3 defines families. Try only the entries from another family, in pool order. An `auto` or `inherit-parent` entry is the parent model, so it never judges. With the default pool, a Claude parent gets `grok-4.7:xhigh`. A parent from any other family gets `claude-opus-5-5:xhigh` first. If the Task tool rejects an entry, use a valid slug of the same model at the closest level or speed from the error message, or move to the next entry when there is none. If no entry from another family runs, try the default pool's entries from another family the same way. If none of those runs, skip the judge, pick the base alone in Phase D, and say so. If you do not know the parent's model, try the pool's real slugs in pool order and say that the family check did not run. The judge sees the rubric and the candidates by path label, scores each criterion, and recommends a base with rationale. It runs in parallel with your reading in Phase D. Don't spawn it while candidates are still writing.
 
 ## Phase D: Pick a base
 

package/skills/check-compiler-errors/SKILL.md

@@ -0,0 +1,23 @@
+---
+name: check-compiler-errors
+description: Run compile and type-check commands and report failures
+---
+
+# Check compiler errors
+
+## Trigger
+
+Compile or type-check failures are blocking local validation or CI.
+
+## Workflow
+
+1. Run the repo's compile and type-check commands.
+2. Summarize errors by file and type.
+3. Fix the highest-confidence issues first.
+4. Re-run checks until clean or blocked.
+
+## Output
+
+- Current compile and type-check status
+- Error summary grouped by file and category
+- Fixes applied and remaining blockers