* docs: add investigation reports for 5 open GitHub issues Comprehensive analysis of issues #543, #544, #545, #555, and #557: - #557: settings.json not generated, module loader error (node/bun mismatch) - #555: Windows hooks not executing, hasIpc always false - #545: formatTool crashes on non-JSON tool_input strings - #544: mem-search skill hint shown incorrectly to Claude Code users - #543: /claude-mem slash command unavailable despite installation Each report includes root cause analysis, affected files, and proposed fixes. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com> * fix(logger): handle non-JSON tool_input in formatTool (#545) Wrap JSON.parse in try-catch to handle raw string inputs (e.g., Bash commands) that aren't valid JSON. Falls back to using the string as-is. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com> * fix(context): update mem-search hint to reference MCP tools (#544) Update hint messages to reference MCP tools (search, get_observations) instead of the deprecated "mem-search skill" terminology. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com> * fix(settings): auto-create settings.json on first load (#557, #543) When settings.json doesn't exist, create it with defaults instead of returning in-memory defaults. Creates parent directory if needed. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com> * fix(hooks): use bun runtime for hooks except smart-install (#557) Change hook commands from node to bun since hooks use bun:sqlite. Keep smart-install.js on node since it bootstraps bun installation. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com> * chore: rebuild plugin scripts * docs: clarify that build artifacts must be committed * fix(docs): update build artifacts directory reference in CLAUDE.md * test: add test coverage for PR #558 fixes - Fix 2 failing tests: update "mem-search skill" → "MCP tools" expectations - Add 56 tests for formatTool() JSON.parse crash fix (Issue #545) - Add 27 tests for settings.json auto-creation (Issue #543) Test coverage includes: - formatTool: JSON parsing, raw strings, objects, null/undefined, all tool types - Settings: file creation, directory creation, schema migration, edge cases 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com> * fix(tests): clean up flaky tests and fix circular dependency Phase 1 of test quality improvements: - Delete 6 harmful/worthless test files that used problematic mock.module() patterns or tested implementation details rather than behavior: - context-builder.test.ts (tested internal implementation) - export-types.test.ts (fragile mock patterns) - smart-install.test.ts (shell script testing antipattern) - session_id_refactor.test.ts (outdated, tested refactoring itself) - validate_sql_update.test.ts (one-time migration validation) - observation-broadcaster.test.ts (excessive mocking) - Fix circular dependency between logger.ts and SettingsDefaultsManager.ts by using late binding pattern - logger now lazily loads settings - Refactor mock.module() to spyOn() in several test files for more maintainable and less brittle tests: - observation-compiler.test.ts - gemini_agent.test.ts - error-handler.test.ts - server.test.ts - response-processor.test.ts All 649 tests pass. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com> * refactor(tests): phase 2 - reduce mock-heavy tests and improve focus - Remove mock-heavy query tests from observation-compiler.test.ts, keep real buildTimeline tests - Convert session_id_usage_validation.test.ts from 477 to 178 lines of focused smoke tests - Remove tests for language built-ins from worker-spawn.test.ts (JSON.parse, array indexing) - Rename logger-coverage.test.ts to logger-usage-standards.test.ts for clarity 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com> * docs(tests): phase 3 - add JSDoc mock justification to test files Document mock usage rationale in 5 test files to improve maintainability: - error-handler.test.ts: Express req/res mocks, logger spies (~11%) - fallback-error-handler.test.ts: Zero mocks, pure function tests - session-cleanup-helper.test.ts: Session fixtures, worker mocks (~19%) - hook-constants.test.ts: process.platform mock for Windows tests (~12%) - session_store.test.ts: Zero mocks, real SQLite :memory: database Part of ongoing effort to document mock justifications per TESTING.md guidelines. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com> * test(integration): phase 5 - add 72 tests for critical coverage gaps Add comprehensive test coverage for previously untested areas: - tests/integration/hook-execution-e2e.test.ts (10 tests) Tests lifecycle hooks execution flow and context propagation - tests/integration/worker-api-endpoints.test.ts (19 tests) Tests all worker service HTTP endpoints without heavy mocking - tests/integration/chroma-vector-sync.test.ts (16 tests) Tests vector embedding synchronization with ChromaDB - tests/utils/tag-stripping.test.ts (27 tests) Tests privacy tag stripping utilities for both <private> and <meta-observation> tags All tests use real implementations where feasible, following the project's testing philosophy of preferring integration-style tests over unit tests with extensive mocking. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com> * context update * docs: add comment linking DEFAULT_DATA_DIR locations Added NOTE comment in logger.ts pointing to the canonical DEFAULT_DATA_DIR in SettingsDefaultsManager.ts. This addresses PR reviewer feedback about the fragility of having the default defined in two places to avoid circular dependencies. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.5 <noreply@anthropic.com>
9.1 KiB
Issue #543 Analysis: /claude-mem Slash Command Not Available Despite Installation
Date: 2026-01-05 Version Analyzed: 8.5.9 Status: Expected Behavior - No such command exists Related Issues: #557 (if it exists), Windows initialization issues
Issue Summary
A user reports that the /claude-mem diagnostics command returns "Unknown slash command: claude-mem" after installing claude-mem v8.5.9 on Windows.
Reported Environment
- Claude-mem version: 8.5.9
- Claude Code version: 2.0.76
- Node.js version: v22.21.0
- Bun version: 1.3.5
- OS: Windows 10.0.26200.7462 (x64)
Reported Plugin Status
- Worker Running: No
- Database Exists: Yes (4.00 KB - minimal/empty database)
- Settings Exist: No
Root Cause Analysis
Finding 1: No /claude-mem Slash Command Exists
Critical Discovery: The /claude-mem diagnostics command does not exist in claude-mem. After extensive codebase analysis:
-
No slash command registration found: The
plugin/commands/directory is empty. Claude-mem does not register any slash commands. -
Skills, not commands: Claude-mem uses Claude Code's skill system, not the command system. Skills are defined in
plugin/skills/:mem-search/- Memory search functionalitytroubleshoot/- Troubleshooting functionalitysearch/- Search operationsclaude-mem-settings/- Settings management
-
Empty skill directories: All skill directories currently contain only empty subdirectories (
operations/,principles/) with no SKILL.md files present in the built plugin. This suggests either:- Skills are dynamically loaded from the worker service
- A build issue where skill files are not being bundled
- Skills were removed or relocated in a recent refactor
Finding 2: How Troubleshooting Actually Works
According to the documentation (docs/public/troubleshooting.mdx):
"Describe any issues you're experiencing to Claude, and the troubleshoot skill will automatically activate to provide diagnosis and fixes."
The troubleshoot skill is designed to be invoked naturally - users describe their problem to Claude, and the skill auto-invokes. There is no /claude-mem diagnostics command.
Finding 3: Settings.json Creation Flow
The settings.json file is not created during installation. It is created:
-
On first worker API call: The
ensureSettingsFile()method inSettingsRoutes.ts(lines 400-413) creates the file with defaults when the settings API is first accessed. -
Worker must be running: Since settings creation is triggered by API calls, the worker service must be running for settings to be created.
-
Lazy initialization pattern: This is intentional - settings are created on-demand with sensible defaults rather than during installation.
Finding 4: Worker Service Not Running
The user reports "Worker Running: No". This is the core issue because:
-
Worker auto-start on SessionStart: The
hooks.jsonshows the worker starts via:{ "type": "command", "command": "bun \"${CLAUDE_PLUGIN_ROOT}/scripts/worker-service.cjs\" start", "timeout": 60 } -
Smart-install runs first: Before worker start,
smart-install.jsruns to ensure Bun and uv are installed. -
Windows-specific issues: The user is on Windows, which has known issues:
- PowerShell escaping problems in
cleanupOrphanedProcesses()(Issue #517) - PATH issues with freshly installed Bun
- Process spawning differences
- PowerShell escaping problems in
Finding 5: Database Size Indicates No Data
The database is 4.00 KB, which is essentially an empty schema:
- No observations recorded
- No sessions created
- Hooks may not have executed successfully
Initialization Flow Analysis
Installation
|
v
First Session Start
|
+---> smart-install.js (ensure Bun + uv)
| |
| +---> May fail silently on Windows (PATH issues)
|
+---> worker-service.cjs start
| |
| +---> Likely failing (worker not running)
|
+---> context-hook.js (requires worker)
| |
| +---> Fails or returns empty (no worker)
|
+---> user-message-hook.js
|
+---> No context injected
Why Skills Directories Are Empty
After investigation, the skill directories in plugin/skills/ are scaffolding structures but appear to have no SKILL.md files in the built plugin. The actual skill functionality may be:
- Served via HTTP API: The Server.ts shows an
/api/instructionsendpoint that loads SKILL.md sections on-demand from../skills/mem-search/ - Built differently: The skills may be bundled into the worker service rather than standalone files
- Documentation discrepancy: The README and docs reference skills that may work differently than traditional Claude Code skill files
Proposed Diagnosis
The user's issue is not that /claude-mem diagnostics doesn't work - that command never existed. The actual issues are:
-
Misunderstanding of troubleshoot functionality: The user expects a slash command but should describe issues naturally to Claude.
-
Worker service failed to start: Root cause for:
- No settings.json created
- Empty database (no observations)
- No context injection working
-
Possible Windows initialization failures:
- Bun may not be in PATH after smart-install
- PowerShell execution policy issues
- Worker spawn failures
Recommended User Resolution
Step 1: Verify Bun Installation
bun --version
If not found, manually install:
powershell -c "irm bun.sh/install.ps1 | iex"
Then restart terminal.
Step 2: Manually Start Worker
cd ~/.claude/plugins/marketplaces/thedotmack
bun plugin/scripts/worker-service.cjs start
Step 3: Verify Worker Health
curl http://localhost:37777/health
Step 4: Create Settings Manually (if needed)
curl http://localhost:37777/api/settings
This will create ~/.claude-mem/settings.json with defaults.
Step 5: For Diagnostics - Natural Language
Instead of /claude-mem diagnostics, describe the issue to Claude:
"I'm having issues with claude-mem. Can you help troubleshoot?"
The troubleshoot skill should auto-invoke if the worker is running.
Proposed Code Improvements
1. Add Diagnostic Slash Command
Create a /claude-mem command for diagnostics. File: plugin/commands/claude-mem.json:
{
"name": "claude-mem",
"description": "Claude-mem diagnostics and status",
"handler": "scripts/diagnostic-command.js"
}
2. Eager Settings Creation
Modify smart-install.js to create settings.json during installation:
const settingsPath = join(homedir(), '.claude-mem', 'settings.json');
if (!existsSync(settingsPath)) {
mkdirSync(join(homedir(), '.claude-mem'), { recursive: true });
writeFileSync(settingsPath, JSON.stringify(getDefaults(), null, 2));
console.log('Created settings.json with defaults');
}
3. Better Windows Error Reporting
Add explicit error messages when worker fails to start on Windows:
if (process.platform === 'win32' && !workerStarted) {
console.error('Worker failed to start on Windows.');
console.error('Please run manually: bun plugin/scripts/worker-service.cjs start');
console.error('And check: https://docs.claude-mem.ai/troubleshooting');
}
4. Health Check Command
Add a simple health check that works without the worker:
// plugin/scripts/health-check.js
const http = require('http');
http.get('http://localhost:37777/health', (res) => {
if (res.statusCode === 200) console.log('Worker: RUNNING');
else console.log('Worker: NOT RESPONDING');
}).on('error', () => console.log('Worker: NOT RUNNING'));
Relationship to Issue #557
If Issue #557 relates to initialization issues, this analysis confirms:
- Settings.json creation is lazy (requires worker)
- Worker auto-start can fail silently on Windows
- Users may have incomplete installations without clear error messages
Files Examined
/plugin/.claude-plugin/plugin.json- Plugin manifest (no commands)/plugin/hooks/hooks.json- Hook definitions/plugin/scripts/smart-install.js- Installation script/plugin/scripts/worker-service.cjs- Worker service/src/services/worker/http/routes/SettingsRoutes.ts- Settings creation/src/shared/SettingsDefaultsManager.ts- Default values/src/shared/paths.ts- Path definitions/docs/public/troubleshooting.mdx- User documentation/docs/public/usage/getting-started.mdx- User guide
Conclusion
The reported issue is a user expectation mismatch combined with a Windows initialization failure:
/claude-mem diagnosticsdoes not exist - users should use natural language to invoke the troubleshoot skill- The worker service failed to start, causing cascading issues (no settings, no context)
- Documentation could be clearer about available commands vs skills
- Windows-specific initialization issues are a known pattern
The fix should include both user documentation improvements and potentially adding a /claude-mem diagnostic command for discoverability.