@a5c-ai/babysitter-opencode 0.1.3 → 0.1.4-staging.06f78f92

package/README.md

@@ -25,6 +25,9 @@ plugins/babysitter-opencode/
     tool-execute-after.js   Post-tool-use hook
   skills/
     babysit/SKILL.md     Core orchestration skill
+    accomplish-status/SKILL.md  Run status reporting for Accomplish
+  accomplish-skills/
+    babysitter/SKILL.md  Bundled skill in Accomplish format
   plugin.json            Plugin manifest
   versions.json          SDK version tracking
 ```
@@ -36,7 +39,7 @@ plugins/babysitter-opencode/
 Install the Babysitter SDK CLI:
 
 ```bash
-npm install -g @a5c-ai/babysitter
+npm install -g @a5c-ai/babysitter-sdk
 ```
 
 ### Method 1: npm global install (recommended)
@@ -54,10 +57,10 @@ To install into a specific workspace:
 babysitter-opencode install --workspace /path/to/project
 ```
 
-### Method 2: Babysitter plugin manager
+### Method 2: Babysitter harness install
 
 ```bash
-babysitter plugin:install babysitter-opencode
+babysitter harness:install-plugin opencode
 ```
 
 ### Method 3: Manual copy
@@ -83,11 +86,34 @@ Or from a specific workspace:
 babysitter-opencode uninstall --workspace /path/to/project
 ```
 
+### Method 4: Accomplish AI (auto-detected)
+
+When [Accomplish](https://github.com/accomplish-ai/accomplish) is installed,
+the installer automatically detects it and copies the plugin into Accomplish's
+OpenCode config directory (`<userDataPath>/opencode/plugins/babysitter/`).
+
+```bash
+# Auto-detects Accomplish during standard install
+npm install -g @a5c-ai/babysitter-opencode
+
+# Or target Accomplish explicitly
+babysitter-opencode install --accomplish
+
+# Install to both standalone OpenCode and Accomplish
+babysitter-opencode install --global --accomplish
+```
+
+Accomplish stores OpenCode config at platform-specific locations:
+- **macOS**: `~/Library/Application Support/Accomplish/opencode/`
+- **Windows**: `%APPDATA%/Accomplish/opencode/`
+- **Linux**: `~/.config/Accomplish/opencode/`
+
 ## CLI Reference
 
 ```
 babysitter-opencode install [--global]            Install plugin globally
 babysitter-opencode install --workspace [path]    Install into workspace
+babysitter-opencode install --accomplish          Install into Accomplish data dir
 babysitter-opencode uninstall [--global]          Uninstall plugin globally
 babysitter-opencode uninstall --workspace [path]  Uninstall from workspace
 babysitter-opencode sync                          Sync command surfaces
@@ -127,6 +153,24 @@ natively provide them:
 - `BABYSITTER_RUNS_DIR` -- Runs directory path
 - `OPENCODE_PLUGIN_ROOT` -- Plugin root directory
 
+### Accomplish Integration
+
+When running inside [Accomplish](https://github.com/accomplish-ai/accomplish), OpenCode is spawned as a subprocess with
+additional environment variables. The babysitter adapter detects these
+automatically:
+
+- `ACCOMPLISH_TASK_ID` -- Accomplish task identifier (primary correlation key)
+- `OPENCODE_CONFIG_DIR` -- Path to Accomplish's OpenCode config directory
+- `OPENCODE_CONFIG` -- Path to the generated `opencode.json`
+
+The adapter writes `accomplishTaskId` into session state metadata for
+correlation. The `accomplish-status` skill writes run status JSON to
+`<OPENCODE_CONFIG_DIR>/run-status/<runId>.json` for file-based IPC.
+
+Breakpoints are resolved conversationally: when a breakpoint is pending, the
+agent uses Accomplish's `ask-user-question` MCP tool to prompt the user, then
+posts the result via `babysitter task:post`.
+
 ### Configuration Environment Variables
 
 | Variable | Default | Description |
@@ -137,6 +181,8 @@ natively provide them:
 | `BABYSITTER_OPENCODE_MARKETPLACE_PATH` | `~/.agents/plugins/marketplace.json` | Marketplace file |
 | `BABYSITTER_SDK_CLI` | (auto-detected) | Path to SDK CLI entry |
 | `BABYSITTER_GLOBAL_STATE_DIR` | `~/.a5c` | Global state directory |
+| `ACCOMPLISH_TASK_ID` | -- | Set by Accomplish when spawning OpenCode |
+| `OPENCODE_CONFIG_DIR` | -- | Accomplish's OpenCode config directory |
 
 ## Verification
 

package/bin/install-shared.cjs

@@ -367,6 +367,76 @@ function ensureGlobalProcessLibrary(packageRoot) {
   );
 }
 
+// ---------------------------------------------------------------------------
+// Accomplish AI detection and paths
+// ---------------------------------------------------------------------------
+
+/**
+ * Returns the Accomplish user data directory for the current platform.
+ * If OPENCODE_CONFIG_DIR is set, returns its parent (the Accomplish data dir).
+ * Otherwise falls back to platform-specific defaults.
+ */
+function getAccomplishDataDir() {
+  if (process.env.OPENCODE_CONFIG_DIR) {
+    return path.resolve(process.env.OPENCODE_CONFIG_DIR, '..');
+  }
+  const home = getUserHome();
+  switch (process.platform) {
+    case 'darwin':
+      return path.join(home, 'Library', 'Application Support', 'Accomplish');
+    case 'win32': {
+      const appData = process.env.APPDATA || process.env.LOCALAPPDATA;
+      return appData
+        ? path.join(appData, 'Accomplish')
+        : path.join(home, 'AppData', 'Roaming', 'Accomplish');
+    }
+    default:
+      return path.join(home, '.config', 'Accomplish');
+  }
+}
+
+/**
+ * Returns true if Accomplish AI appears to be installed or is running.
+ * Checks for:
+ *  - ACCOMPLISH_TASK_ID env var (running inside Accomplish)
+ *  - Accomplish data dir with an opencode/ subdirectory on disk
+ */
+function isAccomplishInstalled() {
+  if (process.env.ACCOMPLISH_TASK_ID) return true;
+  try {
+    const dataDir = getAccomplishDataDir();
+    const openCodeDir = path.join(dataDir, 'opencode');
+    return fs.existsSync(openCodeDir);
+  } catch {
+    return false;
+  }
+}
+
+/**
+ * Returns the OpenCode home directory inside the Accomplish data dir.
+ */
+function getAccomplishOpenCodeHome() {
+  return path.join(getAccomplishDataDir(), 'opencode');
+}
+
+/**
+ * Install the babysitter plugin into Accomplish's OpenCode directory.
+ * Mirrors the standard OpenCode install: copies bundle, writes index.js,
+ * installs skills and hooks.
+ */
+function installAccomplishSurface(packageRoot, accomplishOpenCodeHome) {
+  const pluginRoot = path.join(accomplishOpenCodeHome, 'plugins', PLUGIN_NAME);
+
+  // Copy plugin bundle
+  copyPluginBundle(packageRoot, pluginRoot);
+
+  // Create index.js entry point
+  writeIndexJs(pluginRoot);
+
+  // Install skills and hooks config
+  installOpenCodeSurface(packageRoot, accomplishOpenCodeHome);
+}
+
 // ---------------------------------------------------------------------------
 // OpenCode surface installation
 // ---------------------------------------------------------------------------
@@ -395,10 +465,14 @@ module.exports = {
   copyRecursive,
   ensureGlobalProcessLibrary,
   ensureMarketplaceEntry,
+  getAccomplishDataDir,
+  getAccomplishOpenCodeHome,
   getHomeMarketplacePath,
   getHomePluginRoot,
   getOpenCodeHome,
+  installAccomplishSurface,
   installOpenCodeSurface,
+  isAccomplishInstalled,
   removeManagedHooks,
   removeMarketplaceEntry,
   writeIndexJs,

package/bin/install.cjs

@@ -14,6 +14,7 @@
  *   node install.cjs                     # Install into cwd workspace
  *   node install.cjs --workspace /path   # Install into specified workspace
  *   node install.cjs --global            # Global install (user home)
+ *   node install.cjs --accomplish        # Install only to Accomplish AI data dir
  */
 
 const path = require('path');
@@ -21,10 +22,13 @@ const {
   copyPluginBundle,
   ensureGlobalProcessLibrary,
   ensureMarketplaceEntry,
+  getAccomplishOpenCodeHome,
   getHomeMarketplacePath,
   getHomePluginRoot,
   getOpenCodeHome,
+  installAccomplishSurface,
   installOpenCodeSurface,
+  isAccomplishInstalled,
   writeIndexJs,
 } = require('./install-shared.cjs');
 
@@ -32,6 +36,7 @@ const PACKAGE_ROOT = path.resolve(__dirname, '..');
 
 function parseArgs(argv) {
   let workspace = process.env.OPENCODE_WORKSPACE || process.cwd();
+  let accomplish = false;
   for (let i = 2; i < argv.length; i += 1) {
     const arg = argv[i];
     if (arg === '--workspace') {
@@ -43,47 +48,91 @@ function parseArgs(argv) {
       workspace = null;
       continue;
     }
+    if (arg === '--accomplish') {
+      accomplish = true;
+      continue;
+    }
     throw new Error(`unknown argument: ${arg}`);
   }
-  return { workspace };
+  return { workspace, accomplish };
 }
 
-function main() {
-  const { workspace } = parseArgs(process.argv);
+function installStandardOpenCode(workspace) {
   const openCodeHome = getOpenCodeHome(workspace);
   const pluginRoot = getHomePluginRoot(workspace);
   const marketplacePath = getHomeMarketplacePath(workspace);
 
   console.log(`[babysitter] Installing OpenCode plugin to ${pluginRoot}`);
 
+  // 1. Copy plugin bundle
+  copyPluginBundle(PACKAGE_ROOT, pluginRoot);
+  console.log('[babysitter]   Copied plugin bundle');
+
+  // 2. Write index.js entry point for OpenCode plugin discovery
+  writeIndexJs(pluginRoot);
+  console.log('[babysitter]   Created index.js entry point');
+
+  // 3. Register in marketplace
+  ensureMarketplaceEntry(marketplacePath, pluginRoot);
+  console.log(`[babysitter]   Marketplace: ${marketplacePath}`);
+
+  // 4. Install OpenCode surfaces (skills, hooks config)
+  installOpenCodeSurface(PACKAGE_ROOT, openCodeHome);
+  console.log('[babysitter]   Installed hooks and skills');
+
+  // 5. Bootstrap global process library
+  try {
+    const active = ensureGlobalProcessLibrary(PACKAGE_ROOT);
+    console.log(`[babysitter]   Process library: ${active.binding?.dir || '(default)'}`);
+    if (active.defaultSpec?.cloneDir) {
+      console.log(`[babysitter]   Process library clone: ${active.defaultSpec.cloneDir}`);
+    }
+    console.log(`[babysitter]   Process library state: ${active.stateFile}`);
+  } catch (err) {
+    console.warn(`[babysitter]   Warning: Could not bootstrap process library: ${err.message}`);
+    console.warn('[babysitter]   Run "babysitter process-library:clone" manually if needed.');
+  }
+}
+
+function installToAccomplish() {
+  const accomplishHome = getAccomplishOpenCodeHome();
+  console.log(`[babysitter] Installing plugin to Accomplish AI: ${accomplishHome}`);
+
+  installAccomplishSurface(PACKAGE_ROOT, accomplishHome);
+
+  console.log('[babysitter]   Copied plugin bundle to Accomplish');
+  console.log('[babysitter]   Created index.js entry point for Accomplish');
+  console.log('[babysitter]   Installed hooks and skills for Accomplish');
+  console.log('[babysitter] Accomplish AI installation complete.');
+  console.log('[babysitter] Restart Accomplish to pick up the installed plugin.');
+}
+
+function main() {
+  const { workspace, accomplish } = parseArgs(process.argv);
+
   try {
-    // 1. Copy plugin bundle
-    copyPluginBundle(PACKAGE_ROOT, pluginRoot);
-    console.log('[babysitter]   Copied plugin bundle');
-
-    // 2. Write index.js entry point for OpenCode plugin discovery
-    writeIndexJs(pluginRoot);
-    console.log('[babysitter]   Created index.js entry point');
-
-    // 3. Register in marketplace
-    ensureMarketplaceEntry(marketplacePath, pluginRoot);
-    console.log(`[babysitter]   Marketplace: ${marketplacePath}`);
-
-    // 4. Install OpenCode surfaces (skills, hooks config)
-    installOpenCodeSurface(PACKAGE_ROOT, openCodeHome);
-    console.log('[babysitter]   Installed hooks and skills');
-
-    // 5. Bootstrap global process library
-    try {
-      const active = ensureGlobalProcessLibrary(PACKAGE_ROOT);
-      console.log(`[babysitter]   Process library: ${active.binding?.dir || '(default)'}`);
-      if (active.defaultSpec?.cloneDir) {
-        console.log(`[babysitter]   Process library clone: ${active.defaultSpec.cloneDir}`);
+    // If --accomplish is used without --global/--workspace, install only to Accomplish
+    if (accomplish && workspace !== null) {
+      installToAccomplish();
+      return;
+    }
+
+    // Standard OpenCode install
+    installStandardOpenCode(workspace);
+
+    // Auto-detect Accomplish and install there too (unless --accomplish was explicit)
+    if (!accomplish) {
+      try {
+        if (isAccomplishInstalled()) {
+          console.log('[babysitter] Detected Accomplish AI installation.');
+          installToAccomplish();
+        }
+      } catch (err) {
+        console.warn(`[babysitter]   Warning: Accomplish AI detection failed: ${err.message}`);
       }
-      console.log(`[babysitter]   Process library state: ${active.stateFile}`);
-    } catch (err) {
-      console.warn(`[babysitter]   Warning: Could not bootstrap process library: ${err.message}`);
-      console.warn('[babysitter]   Run "babysitter process-library:clone" manually if needed.');
+    } else {
+      // --accomplish combined with --global: install to both
+      installToAccomplish();
     }
 
     console.log('[babysitter] Installation complete!');

package/commands/doctor.md

@@ -4,9 +4,9 @@ argument-hint: "[run-id] Optional run ID to diagnose. If omitted, uses the most
 allowed-tools: Read, Grep, Write, Task, Bash, Edit, Grep, Glob, WebFetch, WebSearch, Search, AskUserQuestion, TodoWrite, TodoRead, Skill, BashOutput, KillShell, MultiEdit, LS
 ---
 
-You are a diagnostic agent for the babysitter runtime. Your job is to perform a comprehensive health check across 10 areas and produce a structured diagnostic report. Follow each section methodically. Track results as you go and produce the final summary at the end.
+You are a diagnostic agent for the babysitter runtime. Your job is to perform a comprehensive health check across 14 areas and produce a structured diagnostic report. Follow each section methodically. Track results as you go and produce the final summary at the end.
 
-Initialize a results tracker with these 10 checks, all starting as PENDING:
+Initialize a results tracker with these 14 checks, all starting as PENDING:
 1. Run Discovery
 2. Journal Integrity
 3. State Cache Consistency
@@ -17,6 +17,10 @@ Initialize a results tracker with these 10 checks, all starting as PENDING:
 8. Disk Usage
 9. Process Validation
 10. Hook Execution Health
+11. Session-ID Provenance
+12. Ancestor Liveness
+13. Concurrent Session Detection
+14. Windows Ancestor-Walk Strategy
 
 ---
 
@@ -177,13 +181,13 @@ Mark as PASS if no issues. Mark as WARN if runaway loops or stale sessions detec
 **Goal:** Analyze babysitter log files for errors, warnings, and stop hook decisions.
 
 Read the last 50 lines of each of these log files (if they exist):
-- `$CLAUDE_PLUGIN_ROOT/.a5c/logs/hooks.log`
-- `$CLAUDE_PLUGIN_ROOT/.a5c/logs/babysitter-stop-hook.log`
-- `$CLAUDE_PLUGIN_ROOT/.a5c/logs/babysitter-stop-hook-stderr.log`
-- `$CLAUDE_PLUGIN_ROOT/.a5c/logs/babysitter-session-start-hook.log`
-- `$CLAUDE_PLUGIN_ROOT/.a5c/logs/babysitter-session-start-hook-stderr.log`
-- `$CLAUDE_PLUGIN_ROOT/.a5c/logs/babysitter.log`
-- `$HOME/.a5c/logs/` and relevant logs and run/session specific logs there
+- `${BABYSITTER_LOG_DIR:-$HOME/.a5c/logs}/hooks.log`
+- `${BABYSITTER_LOG_DIR:-$HOME/.a5c/logs}/babysitter-stop-hook.log`
+- `${BABYSITTER_LOG_DIR:-$HOME/.a5c/logs}/babysitter-stop-hook-stderr.log`
+- `${BABYSITTER_LOG_DIR:-$HOME/.a5c/logs}/babysitter-session-start-hook.log`
+- `${BABYSITTER_LOG_DIR:-$HOME/.a5c/logs}/babysitter-session-start-hook-stderr.log`
+- `${BABYSITTER_LOG_DIR:-$HOME/.a5c/logs}/babysitter.log`
+- `${BABYSITTER_LOG_DIR:-$HOME/.a5c/logs}/` and relevant run/session specific logs there
 
 
 For each log file:
@@ -285,7 +289,7 @@ The hooks delegate to the `babysitter` CLI. Check if it is available:
 Check whether the stop hook has actually been invoked during this run's lifetime:
 
 **From log files:**
-- Read `$CLAUDE_PLUGIN_ROOT/.a5c/logs/babysitter-stop-hook.log` (if it exists).
+- Read `${BABYSITTER_LOG_DIR:-$HOME/.a5c/logs}/babysitter-stop-hook.log` (if it exists).
 - Count the number of "Hook script invoked" lines. This is the total invocation count.
 - Count the number of "CLI exit code=" lines and extract exit codes.
 - If the log file does not exist or has zero invocations, the stop hook has NOT been running.
@@ -297,7 +301,7 @@ Check whether the stop hook has actually been invoked during this run's lifetime
 - If no STOP_HOOK_INVOKED events exist in the journal, note that the stop hook has not recorded any decisions for this run.
 
 **From stderr:**
-- Read `$CLAUDE_PLUGIN_ROOT/.a5c/logs/babysitter-stop-hook-stderr.log`.
+- Read `${BABYSITTER_LOG_DIR:-$HOME/.a5c/logs}/babysitter-stop-hook-stderr.log`.
 - If it contains error output, display it and diagnose:
   - "command not found" or exit code 127 → CLI not installed (see 10c)
   - "MODULE_NOT_FOUND" or "Cannot find module" → SDK package corrupted or not built
@@ -350,9 +354,65 @@ Mark as FAIL if:
 
 ---
 
+## 11. Session-ID Provenance
+
+**Goal:** Verify how the current babysitter session ID was resolved and flag stale or shadowed values.
+
+- Invoke: `npx babysitter session:whoami --json`
+- Parse the output and inspect the `resolvedFrom` field. Classify as follows:
+  - `resolvedFrom: "pid-marker"` → mark as PASS ("Session ID derives from the live Claude Code ancestor process -- authoritative").
+  - `resolvedFrom: "env-file"` → mark as PASS with a note ("CLAUDE_ENV_FILE was used; typically healthy").
+  - `resolvedFrom: "env-var"` → mark as WARN ("`BABYSITTER_SESSION_ID` is set without a corroborating PID marker. Likely stale from a prior Claude Code session -- see GitHub issue #130").
+    - Remediation: run `babysitter session:cleanup` and start a fresh Claude Code session, or `unset BABYSITTER_SESSION_ID` before invoking babysitter.
+  - `resolvedFrom: "none"` → mark as ERROR ("No session ID resolvable. Either no session-start hook fired, or the ancestor walk failed").
+
+**Env-var shadow check:**
+- Independently inspect `envVarPresent` and `envVarMatches` in the output.
+- If `envVarPresent && !envVarMatches`, mark as WARN ("`BABYSITTER_SESSION_ID` in env does not match the resolved session ID; a stale value is shadowing the authoritative one. Unset the env var").
+
+---
+
+## 12. Ancestor Liveness
+
+**Goal:** Confirm the PID marker references a live Claude Code process.
+
+- Reuse the `session:whoami --json` output from check 11.
+- Inspect the `ancestorAlive` field.
+- If `ancestorAlive === false`, mark as ERROR ("The PID marker references a dead Claude Code process").
+  - Remediation: `babysitter session:cleanup`.
+- Otherwise mark as PASS.
+
+---
+
+## 13. Concurrent Session Detection
+
+**Goal:** Surface multiple live harness sessions that may compete for the same session ID.
+
+- Enumerate files in `~/.a5c/` matching the pattern `current-session-*-pid-*`.
+- Count markers per harness (derived from the filename).
+- If more than one live marker exists for the same harness, mark as INFO ("Multiple live Claude Code / harness sessions detected; ensure each shell scopes `BABYSITTER_SESSION_ID` appropriately -- the PID marker handles this automatically").
+- Otherwise mark as PASS.
+
+---
+
+## 14. Windows Ancestor-Walk Strategy
+
+**Goal:** Verify the ancestor-walk strategy works on Windows, where `wmic` is no longer guaranteed to be present.
+
+- Only run this check when `process.platform === 'win32'`. On other platforms, mark as PASS ("Not applicable -- non-Windows platform").
+- Attempt the ancestor walk by invoking `npx babysitter session:whoami --json` (reuse output from check 11 if available).
+- If resolution succeeded (any `resolvedFrom` other than `none`), mark as PASS.
+- If `resolvedFrom: "none"` on Windows:
+  - Test `wmic` availability: `where wmic` via shell.
+  - If absent, document that Windows 11 24H2 removed `wmic`; the fallback PowerShell CIM path should handle this.
+  - If the PowerShell ancestor walk also failed, mark as ERROR with remediation: ensure PowerShell is available (`powershell -NoProfile -Command "Get-CimInstance Win32_Process -Filter ProcessId=$PID"` should work).
+- If the cascade works but is slow (>5s on first probe), add an INFO note on first-probe latency.
+
+---
+
 ## Final Report
 
-After completing all 10 checks, produce the diagnostic report in this format:
+After completing all 14 checks, produce the diagnostic report in this format:
 
 ```
 ============================================
@@ -379,6 +439,10 @@ OVERALL HEALTH: <HEALTHY | WARNING | CRITICAL>
 | 8  | Disk Usage               | <status> |
 | 9  | Process Validation       | <status> |
 | 10 | Hook Execution Health    | <status> |
+| 11 | Session-ID Provenance    | <status> |
+| 12 | Ancestor Liveness        | <status> |
+| 13 | Concurrent Session Detection | <status> |
+| 14 | Windows Ancestor-Walk Strategy | <status> |
 
 --------------------------------------------
   ISSUES & RECOMMENDATIONS
@@ -392,9 +456,9 @@ OVERALL HEALTH: <HEALTHY | WARNING | CRITICAL>
 ```
 
 **Overall health determination:**
-- **HEALTHY**: All 10 checks are PASS.
-- **WARNING**: At least one check is WARN but none are FAIL.
-- **CRITICAL**: At least one check is FAIL.
+- **HEALTHY**: All 14 checks are PASS (INFO notes are acceptable).
+- **WARNING**: At least one check is WARN but none are FAIL or ERROR.
+- **CRITICAL**: At least one check is FAIL or ERROR.
 
 Present the full detailed findings for each check BEFORE the summary table, so the user can see the evidence. End with the summary table and recommendations. Also, create a single HTML report file with all the findings that uses the arwes UI framework and open it for the user in the browser.
 
@@ -424,3 +488,25 @@ After diagnosing issues, prompt the user to report or fix what was found -- they
 Example prompt after diagnosis:
 
 > "Diagnosis found a stale lock -- process 12847 crashed without cleanup. This is a known edge case in the orchestration loop. Even if you don't want to fix it yourself, reporting it helps: run `/babysitter:contrib bug report: orchestration loop doesn't release lock on unhandled rejection` to open an issue."
+
+---
+
+## Self-Heal Suggestions
+
+If any of checks 11-14 surface issues (stale env vars, dead ancestor PIDs, shadowed session IDs, or Windows ancestor-walk failures), suggest the following remediation sequence, in order. Present it as an actionable block:
+
+```bash
+# 1. Cleanup dead markers and orphaned state files
+babysitter session:cleanup --dry-run   # preview
+babysitter session:cleanup             # apply
+
+# 2. Unset a stale env var
+unset BABYSITTER_SESSION_ID
+
+# 3. Re-bind a run explicitly if needed
+babysitter session:resume --session-id <fresh-id> --state-dir ~/.a5c --run-id <runId> --runs-dir .a5c/runs
+
+# 4. Start a fresh Claude Code session (closes and reopens the session)
+```
+
+Run steps 1 and 2 first; re-run `/babysitter:doctor` after each step to confirm the session-provenance checks return to PASS. Step 3 is only needed when a specific run must be re-bound to the fresh session. If the issue persists after step 4, escalate via `/debug` or `/babysitter:contrib`.

package/hooks/session-created.js

@@ -16,12 +16,13 @@
 
 const { execSync, execFileSync } = require("child_process");
 const { readFileSync, mkdirSync, appendFileSync, existsSync, writeFileSync } = require("fs");
+const os = require("os");
 const path = require("path");
 const crypto = require("crypto");
 
 const PLUGIN_ROOT = process.env.OPENCODE_PLUGIN_ROOT || path.resolve(__dirname, "..");
 const STATE_DIR = process.env.BABYSITTER_STATE_DIR || path.join(process.cwd(), ".a5c");
-const LOG_DIR = process.env.BABYSITTER_LOG_DIR || path.join(PLUGIN_ROOT, ".a5c", "logs");
+const LOG_DIR = process.env.BABYSITTER_LOG_DIR || path.join(os.homedir(), ".a5c", "logs");
 const LOG_FILE = path.join(LOG_DIR, "babysitter-session-created-hook.log");
 
 // ---------------------------------------------------------------------------

package/hooks/session-idle.js

@@ -20,11 +20,12 @@
 
 const { execSync } = require("child_process");
 const { readFileSync, mkdirSync, appendFileSync } = require("fs");
+const os = require("os");
 const path = require("path");
 
 const PLUGIN_ROOT = process.env.OPENCODE_PLUGIN_ROOT || path.resolve(__dirname, "..");
 const STATE_DIR = process.env.BABYSITTER_STATE_DIR || path.join(process.cwd(), ".a5c");
-const LOG_DIR = process.env.BABYSITTER_LOG_DIR || path.join(PLUGIN_ROOT, ".a5c", "logs");
+const LOG_DIR = process.env.BABYSITTER_LOG_DIR || path.join(os.homedir(), ".a5c", "logs");
 const LOG_FILE = path.join(LOG_DIR, "babysitter-session-idle-hook.log");
 
 function ensureDir(dir) {

package/hooks/shell-env.js

@@ -18,13 +18,14 @@
 "use strict";
 
 const { readFileSync, mkdirSync, appendFileSync, existsSync } = require("fs");
+const os = require("os");
 const path = require("path");
 const crypto = require("crypto");
 
 const PLUGIN_ROOT = process.env.OPENCODE_PLUGIN_ROOT || path.resolve(__dirname, "..");
 const STATE_DIR = process.env.BABYSITTER_STATE_DIR || path.join(process.cwd(), ".a5c");
 const RUNS_DIR = process.env.BABYSITTER_RUNS_DIR || path.join(STATE_DIR, "runs");
-const LOG_DIR = process.env.BABYSITTER_LOG_DIR || path.join(PLUGIN_ROOT, ".a5c", "logs");
+const LOG_DIR = process.env.BABYSITTER_LOG_DIR || path.join(os.homedir(), ".a5c", "logs");
 const LOG_FILE = path.join(LOG_DIR, "babysitter-shell-env-hook.log");
 
 function ensureDir(dir) {

package/hooks/tool-execute-after.js

@@ -20,11 +20,12 @@
 
 const { execSync } = require("child_process");
 const { readFileSync, mkdirSync, appendFileSync } = require("fs");
+const os = require("os");
 const path = require("path");
 
 const PLUGIN_ROOT = process.env.OPENCODE_PLUGIN_ROOT || path.resolve(__dirname, "..");
 const STATE_DIR = process.env.BABYSITTER_STATE_DIR || path.join(process.cwd(), ".a5c");
-const LOG_DIR = process.env.BABYSITTER_LOG_DIR || path.join(PLUGIN_ROOT, ".a5c", "logs");
+const LOG_DIR = process.env.BABYSITTER_LOG_DIR || path.join(os.homedir(), ".a5c", "logs");
 const LOG_FILE = path.join(LOG_DIR, "babysitter-tool-after-hook.log");
 
 function ensureDir(dir) {

package/hooks/tool-execute-before.js

@@ -20,11 +20,12 @@
 
 const { execSync } = require("child_process");
 const { readFileSync, mkdirSync, appendFileSync } = require("fs");
+const os = require("os");
 const path = require("path");
 
 const PLUGIN_ROOT = process.env.OPENCODE_PLUGIN_ROOT || path.resolve(__dirname, "..");
 const STATE_DIR = process.env.BABYSITTER_STATE_DIR || path.join(process.cwd(), ".a5c");
-const LOG_DIR = process.env.BABYSITTER_LOG_DIR || path.join(PLUGIN_ROOT, ".a5c", "logs");
+const LOG_DIR = process.env.BABYSITTER_LOG_DIR || path.join(os.homedir(), ".a5c", "logs");
 const LOG_FILE = path.join(LOG_DIR, "babysitter-tool-before-hook.log");
 
 function ensureDir(dir) {

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@a5c-ai/babysitter-opencode",
-  "version": "0.1.3",
+  "version": "0.1.4-staging.06f78f92",
   "description": "Babysitter orchestration plugin for OpenCode with SDK-managed process-library bootstrapping and in-turn iteration model",
   "scripts": {
     "test": "node test/integration.test.js",
@@ -41,6 +41,6 @@
   },
   "homepage": "https://github.com/a5c-ai/babysitter/tree/main/plugins/babysitter-opencode#readme",
   "dependencies": {
-    "@a5c-ai/babysitter-sdk": "0.0.187"
+    "@a5c-ai/babysitter-sdk": "0.0.188-staging.06f78f92"
   }
 }

package/skills/accomplish-status/SKILL.md

@@ -0,0 +1,114 @@
+---
+name: accomplish-status
+description: Report babysitter orchestration run status to Accomplish AI desktop app via file-based IPC
+command: /accomplish-status
+---
+
+# accomplish-status
+
+Report babysitter run status to the [Accomplish AI](https://github.com/accomplish-ai/accomplish) desktop app via file-based IPC.
+
+## Overview
+
+This skill writes run status JSON to a well-known directory. External consumers (such as a future Accomplish daemon integration or other tooling) can watch this directory for changes to track orchestration progress, pending breakpoints, and completion state.
+
+## Status File Location
+
+Status files are written to:
+
+```
+<OPENCODE_CONFIG_DIR>/run-status/<runId>.json
+```
+
+Where `OPENCODE_CONFIG_DIR` is resolved from the `OPENCODE_CONFIG_DIR` environment variable (typically `~/.config/opencode` or platform equivalent).
+
+## Dependencies
+
+### Babysitter SDK and CLI
+
+Read the SDK version from `versions.json` to ensure version compatibility:
+
+```bash
+SDK_VERSION=$(node -e "try{const fs=require('fs');const probes=['./plugins/babysitter-opencode/versions.json','./node_modules/@a5c-ai/babysitter-opencode/versions.json'];for(const probe of probes){if(fs.existsSync(probe)){console.log(JSON.parse(fs.readFileSync(probe,'utf8')).sdkVersion||'latest');process.exit(0)}}console.log('latest')}catch{console.log('latest')}")
+CLI="npx -y @a5c-ai/babysitter-sdk@$SDK_VERSION"
+```
+
+## Instructions
+
+### 1. Resolve the status directory
+
+```bash
+STATUS_DIR="${OPENCODE_CONFIG_DIR:-$HOME/.config/opencode}/run-status"
+mkdir -p "$STATUS_DIR"
+```
+
+### 2. Get run status from babysitter
+
+```bash
+babysitter run:status .a5c/runs/<runId> --json
+```
+
+### 3. Write the status file
+
+Transform the run status output into the Accomplish status format and write it to `$STATUS_DIR/<runId>.json`.
+
+Use the `ACCOMPLISH_TASK_ID` environment variable to correlate the babysitter run with the originating Accomplish task. This variable is set automatically by Accomplish when it spawns the agent session.
+
+```bash
+ACCOMPLISH_TASK_ID="${ACCOMPLISH_TASK_ID:-}"
+```
+
+### 4. Status file format
+
+Write the following JSON structure to `$STATUS_DIR/<runId>.json`:
+
+```json
+{
+  "runId": "abc123-def456",
+  "processId": "my-process",
+  "status": "running",
+  "currentPhase": "execute",
+  "phases": [
+    { "name": "plan", "status": "completed" },
+    { "name": "execute", "status": "running" },
+    { "name": "verify", "status": "pending" }
+  ],
+  "pendingEffects": [
+    {
+      "effectId": "eff-001",
+      "kind": "breakpoint",
+      "title": "Approve implementation plan"
+    }
+  ],
+  "accomplishTaskId": "accomplish-task-789",
+  "lastUpdatedAt": "2026-04-04T12:00:00.000Z"
+}
+```
+
+#### Field reference
+
+| Field | Type | Description |
+|-------|------|-------------|
+| `runId` | string | Babysitter run identifier |
+| `processId` | string | Process definition identifier |
+| `status` | enum | One of: `created`, `running`, `waiting`, `completed`, `failed` |
+| `currentPhase` | string | Name of the currently active phase |
+| `phases` | array | Ordered list of phases with individual status (`pending`, `running`, `completed`, `failed`) |
+| `pendingEffects` | array | Currently pending effects (tasks, breakpoints, sleeps) awaiting resolution |
+| `accomplishTaskId` | string | Correlation ID from `ACCOMPLISH_TASK_ID` env var; links this run to the Accomplish UI task |
+| `lastUpdatedAt` | string | ISO 8601 timestamp of the last status update |
+
+### 5. When to update
+
+Write or update the status file:
+
+- Immediately after `babysitter run:create` (status: `created`)
+- After each `babysitter run:iterate` call (status: `running` or `waiting`)
+- When breakpoints are requested (status: `waiting`, include breakpoint in `pendingEffects`)
+- After `babysitter task:post` resolves an effect (status: `running`)
+- On run completion (status: `completed`)
+- On run failure (status: `failed`)
+
+### 6. Cleanup
+
+When a run reaches a terminal state (`completed` or `failed`), the status file may be left in place for external consumers to read the final state. Stale status files can be cleaned up manually or by external tooling.

package/skills/doctor/SKILL.md

@@ -5,9 +5,9 @@ description: Diagnose babysitter run health - journal integrity, state cache, ef
 
 # doctor
 
-You are a diagnostic agent for the babysitter runtime. Your job is to perform a comprehensive health check across 10 areas and produce a structured diagnostic report. Follow each section methodically. Track results as you go and produce the final summary at the end.
+You are a diagnostic agent for the babysitter runtime. Your job is to perform a comprehensive health check across 14 areas and produce a structured diagnostic report. Follow each section methodically. Track results as you go and produce the final summary at the end.
 
-Initialize a results tracker with these 10 checks, all starting as PENDING:
+Initialize a results tracker with these 14 checks, all starting as PENDING:
 1. Run Discovery
 2. Journal Integrity
 3. State Cache Consistency
@@ -18,6 +18,10 @@ Initialize a results tracker with these 10 checks, all starting as PENDING:
 8. Disk Usage
 9. Process Validation
 10. Hook Execution Health
+11. Session-ID Provenance
+12. Ancestor Liveness
+13. Concurrent Session Detection
+14. Windows Ancestor-Walk Strategy
 
 ---
 
@@ -178,13 +182,13 @@ Mark as PASS if no issues. Mark as WARN if runaway loops or stale sessions detec
 **Goal:** Analyze babysitter log files for errors, warnings, and stop hook decisions.
 
 Read the last 50 lines of each of these log files (if they exist):
-- `$CLAUDE_PLUGIN_ROOT/.a5c/logs/hooks.log`
-- `$CLAUDE_PLUGIN_ROOT/.a5c/logs/babysitter-stop-hook.log`
-- `$CLAUDE_PLUGIN_ROOT/.a5c/logs/babysitter-stop-hook-stderr.log`
-- `$CLAUDE_PLUGIN_ROOT/.a5c/logs/babysitter-session-start-hook.log`
-- `$CLAUDE_PLUGIN_ROOT/.a5c/logs/babysitter-session-start-hook-stderr.log`
-- `$CLAUDE_PLUGIN_ROOT/.a5c/logs/babysitter.log`
-- `$HOME/.a5c/logs/` and relevant logs and run/session specific logs there
+- `${BABYSITTER_LOG_DIR:-$HOME/.a5c/logs}/hooks.log`
+- `${BABYSITTER_LOG_DIR:-$HOME/.a5c/logs}/babysitter-stop-hook.log`
+- `${BABYSITTER_LOG_DIR:-$HOME/.a5c/logs}/babysitter-stop-hook-stderr.log`
+- `${BABYSITTER_LOG_DIR:-$HOME/.a5c/logs}/babysitter-session-start-hook.log`
+- `${BABYSITTER_LOG_DIR:-$HOME/.a5c/logs}/babysitter-session-start-hook-stderr.log`
+- `${BABYSITTER_LOG_DIR:-$HOME/.a5c/logs}/babysitter.log`
+- `${BABYSITTER_LOG_DIR:-$HOME/.a5c/logs}/` and relevant run/session specific logs there
 
 
 For each log file:
@@ -286,7 +290,7 @@ The hooks delegate to the `babysitter` CLI. Check if it is available:
 Check whether the stop hook has actually been invoked during this run's lifetime:
 
 **From log files:**
-- Read `$CLAUDE_PLUGIN_ROOT/.a5c/logs/babysitter-stop-hook.log` (if it exists).
+- Read `${BABYSITTER_LOG_DIR:-$HOME/.a5c/logs}/babysitter-stop-hook.log` (if it exists).
 - Count the number of "Hook script invoked" lines. This is the total invocation count.
 - Count the number of "CLI exit code=" lines and extract exit codes.
 - If the log file does not exist or has zero invocations, the stop hook has NOT been running.
@@ -298,7 +302,7 @@ Check whether the stop hook has actually been invoked during this run's lifetime
 - If no STOP_HOOK_INVOKED events exist in the journal, note that the stop hook has not recorded any decisions for this run.
 
 **From stderr:**
-- Read `$CLAUDE_PLUGIN_ROOT/.a5c/logs/babysitter-stop-hook-stderr.log`.
+- Read `${BABYSITTER_LOG_DIR:-$HOME/.a5c/logs}/babysitter-stop-hook-stderr.log`.
 - If it contains error output, display it and diagnose:
   - "command not found" or exit code 127 → CLI not installed (see 10c)
   - "MODULE_NOT_FOUND" or "Cannot find module" → SDK package corrupted or not built
@@ -351,9 +355,65 @@ Mark as FAIL if:
 
 ---
 
+## 11. Session-ID Provenance
+
+**Goal:** Verify how the current babysitter session ID was resolved and flag stale or shadowed values.
+
+- Invoke: `npx babysitter session:whoami --json`
+- Parse the output and inspect the `resolvedFrom` field. Classify as follows:
+  - `resolvedFrom: "pid-marker"` → mark as PASS ("Session ID derives from the live Claude Code ancestor process -- authoritative").
+  - `resolvedFrom: "env-file"` → mark as PASS with a note ("CLAUDE_ENV_FILE was used; typically healthy").
+  - `resolvedFrom: "env-var"` → mark as WARN ("`BABYSITTER_SESSION_ID` is set without a corroborating PID marker. Likely stale from a prior Claude Code session -- see GitHub issue #130").
+    - Remediation: run `babysitter session:cleanup` and start a fresh Claude Code session, or `unset BABYSITTER_SESSION_ID` before invoking babysitter.
+  - `resolvedFrom: "none"` → mark as ERROR ("No session ID resolvable. Either no session-start hook fired, or the ancestor walk failed").
+
+**Env-var shadow check:**
+- Independently inspect `envVarPresent` and `envVarMatches` in the output.
+- If `envVarPresent && !envVarMatches`, mark as WARN ("`BABYSITTER_SESSION_ID` in env does not match the resolved session ID; a stale value is shadowing the authoritative one. Unset the env var").
+
+---
+
+## 12. Ancestor Liveness
+
+**Goal:** Confirm the PID marker references a live Claude Code process.
+
+- Reuse the `session:whoami --json` output from check 11.
+- Inspect the `ancestorAlive` field.
+- If `ancestorAlive === false`, mark as ERROR ("The PID marker references a dead Claude Code process").
+  - Remediation: `babysitter session:cleanup`.
+- Otherwise mark as PASS.
+
+---
+
+## 13. Concurrent Session Detection
+
+**Goal:** Surface multiple live harness sessions that may compete for the same session ID.
+
+- Enumerate files in `~/.a5c/` matching the pattern `current-session-*-pid-*`.
+- Count markers per harness (derived from the filename).
+- If more than one live marker exists for the same harness, mark as INFO ("Multiple live Claude Code / harness sessions detected; ensure each shell scopes `BABYSITTER_SESSION_ID` appropriately -- the PID marker handles this automatically").
+- Otherwise mark as PASS.
+
+---
+
+## 14. Windows Ancestor-Walk Strategy
+
+**Goal:** Verify the ancestor-walk strategy works on Windows, where `wmic` is no longer guaranteed to be present.
+
+- Only run this check when `process.platform === 'win32'`. On other platforms, mark as PASS ("Not applicable -- non-Windows platform").
+- Attempt the ancestor walk by invoking `npx babysitter session:whoami --json` (reuse output from check 11 if available).
+- If resolution succeeded (any `resolvedFrom` other than `none`), mark as PASS.
+- If `resolvedFrom: "none"` on Windows:
+  - Test `wmic` availability: `where wmic` via shell.
+  - If absent, document that Windows 11 24H2 removed `wmic`; the fallback PowerShell CIM path should handle this.
+  - If the PowerShell ancestor walk also failed, mark as ERROR with remediation: ensure PowerShell is available (`powershell -NoProfile -Command "Get-CimInstance Win32_Process -Filter ProcessId=$PID"` should work).
+- If the cascade works but is slow (>5s on first probe), add an INFO note on first-probe latency.
+
+---
+
 ## Final Report
 
-After completing all 10 checks, produce the diagnostic report in this format:
+After completing all 14 checks, produce the diagnostic report in this format:
 
 ```
 ============================================
@@ -380,6 +440,10 @@ OVERALL HEALTH: <HEALTHY | WARNING | CRITICAL>
 | 8  | Disk Usage               | <status> |
 | 9  | Process Validation       | <status> |
 | 10 | Hook Execution Health    | <status> |
+| 11 | Session-ID Provenance    | <status> |
+| 12 | Ancestor Liveness        | <status> |
+| 13 | Concurrent Session Detection | <status> |
+| 14 | Windows Ancestor-Walk Strategy | <status> |
 
 --------------------------------------------
   ISSUES & RECOMMENDATIONS
@@ -393,9 +457,9 @@ OVERALL HEALTH: <HEALTHY | WARNING | CRITICAL>
 ```
 
 **Overall health determination:**
-- **HEALTHY**: All 10 checks are PASS.
-- **WARNING**: At least one check is WARN but none are FAIL.
-- **CRITICAL**: At least one check is FAIL.
+- **HEALTHY**: All 14 checks are PASS (INFO notes are acceptable).
+- **WARNING**: At least one check is WARN but none are FAIL or ERROR.
+- **CRITICAL**: At least one check is FAIL or ERROR.
 
 Present the full detailed findings for each check BEFORE the summary table, so the user can see the evidence. End with the summary table and recommendations. Also, create a single HTML report file with all the findings that uses the arwes UI framework and open it for the user in the browser.
 
@@ -425,3 +489,25 @@ After diagnosing issues, prompt the user to report or fix what was found -- they
 Example prompt after diagnosis:
 
 > "Diagnosis found a stale lock -- process 12847 crashed without cleanup. This is a known edge case in the orchestration loop. Even if you don't want to fix it yourself, reporting it helps: run `/babysitter:contrib bug report: orchestration loop doesn't release lock on unhandled rejection` to open an issue."
+
+---
+
+## Self-Heal Suggestions
+
+If any of checks 11-14 surface issues (stale env vars, dead ancestor PIDs, shadowed session IDs, or Windows ancestor-walk failures), suggest the following remediation sequence, in order. Present it as an actionable block:
+
+```bash
+# 1. Cleanup dead markers and orphaned state files
+babysitter session:cleanup --dry-run   # preview
+babysitter session:cleanup             # apply
+
+# 2. Unset a stale env var
+unset BABYSITTER_SESSION_ID
+
+# 3. Re-bind a run explicitly if needed
+babysitter session:resume --session-id <fresh-id> --state-dir ~/.a5c --run-id <runId> --runs-dir .a5c/runs
+
+# 4. Start a fresh Claude Code session (closes and reopens the session)
+```
+
+Run steps 1 and 2 first; re-run `/babysitter:doctor` after each step to confirm the session-provenance checks return to PASS. Step 3 is only needed when a specific run must be re-bound to the fresh session. If the issue persists after step 4, escalate via `/debug` or `/babysitter:contrib`.

package/versions.json

@@ -1,4 +1,4 @@
 {
-  "sdkVersion": "0.0.187",
+  "sdkVersion": "0.0.188-staging.06f78f92",
   "pluginVersion": "0.1.0"
 }