Skip to content

docs: add comprehensive troubleshooting guide (fixes #326)#373

Merged
affaan-m merged 2 commits intoaffaan-m:mainfrom
pangerlkr:patch-10
Mar 11, 2026
Merged

docs: add comprehensive troubleshooting guide (fixes #326)#373
affaan-m merged 2 commits intoaffaan-m:mainfrom
pangerlkr:patch-10

Conversation

@pangerlkr
Copy link
Contributor

@pangerlkr pangerlkr commented Mar 10, 2026

Added a comprehensive troubleshooting guide for the Everything Claude Code (ECC) plugin, covering common issues, symptoms, causes, and solutions.

Description

Added a comprehensive troubleshooting guide for the Everything Claude Code (ECC) plugin, covering common issues, symptoms, causes, and solutions.

Description

Adds TROUBLESHOOTING.md - a comprehensive reference guide for resolving common issues with the ECC plugin.

Type of Change

  • docs: Documentation

What's Included

The troubleshooting guide covers:

1. Memory & Context Issues

  • Context window overflow patterns
  • Memory persistence failures
  • Observation file management

2. Agent Harness Failures

  • Agent not found errors
  • Workflow execution hangs
  • Tool use permission errors

3. Hook & Workflow Errors

4. Installation & Setup

  • Plugin loading issues
  • Package manager detection failures
  • Marketplace vs manual install troubleshooting

5. Performance Issues

  • Slow response times
  • High CPU usage
  • Observation file size management

6. Common Error Messages

  • EACCES permission denied
  • MODULE_NOT_FOUND
  • spawn UNKNOWN (Windows line endings)

Resolves

Closes #326

Testing

Additional Notes

This guide will help users:

  • Self-diagnose common issues
  • Reduce duplicate GitHub issues
  • Understand plugin architecture and file locations
  • Debug hook/agent/workflow failures## Type of Change
  • fix: Bug fix
  • feat: New feature
  • refactor: Code refactoring
  • docs: Documentation
  • test: Tests
  • chore: Maintenance/tooling
  • ci: CI/CD changes

Checklist

  • Tests pass locally (node tests/run-all.js)
  • Validation scripts pass
  • Follows conventional commits format
  • Updated relevant documentation

Summary by cubic

Adds TROUBLESHOOTING.md: a guide to quickly diagnose and fix ECC plugin issues, now with safer steps (backup caches/observations, warn on lock‑file removal) and updated commands covering memory/context limits, agent harness failures, hooks/workflows, install/setup, performance, and common errors; fixes #326.

Written for commit b94fdb6. Summary will update on new commits.

Summary by CodeRabbit

  • Documentation
    • Added a comprehensive TROUBLESHOOTING guide for the ECC plugin covering memory & context issues, agent harness failures, hook & workflow errors, installation & setup problems, performance concerns, and common error messages. Includes symptoms, probable causes, step-by-step remediation, diagnostic commands, verification steps, and guidance for collecting info and seeking help.

Added a comprehensive troubleshooting guide for the Everything Claude Code (ECC) plugin, covering common issues, symptoms, causes, and solutions.
Copilot AI review requested due to automatic review settings March 10, 2026 11:27
@coderabbitai
Copy link
Contributor

coderabbitai bot commented Mar 10, 2026

Caution

Review failed

The pull request is closed.

ℹ️ Recent review info
⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro

Run ID: 82351687-ac95-4ebb-b5e0-3d49e4bb7bd9

📥 Commits

Reviewing files that changed from the base of the PR and between 1c6e401 and b94fdb6.

📒 Files selected for processing (1)
  • TROUBLESHOOTING.md

📝 Walkthrough

Walkthrough

Adds a new TROUBLESHOOTING.md documentation file for the Everything Claude Code (ECC) plugin containing structured troubleshooting guidance for memory & context, agent harnesses, hooks/workflows, installation, performance, common errors, and getting help (symptoms, causes, and remediation steps).

Changes

Cohort / File(s) Summary
Troubleshooting Documentation
TROUBLESHOOTING.md
New comprehensive troubleshooting guide (document adds ~422 lines) covering Memory & Context Issues, Agent Harness Failures, Hook & Workflow Errors, Installation & Setup, Performance Issues, Common Error Messages, Getting Help, and Related Documentation; includes symptoms, causes, and shell commands for remediation.

Estimated code review effort

🎯 1 (Trivial) | ⏱️ ~3 minutes

Poem

🐰 I dug a doc for fixes bright and keen,
I hopped through errors, found what's unseen.
When contexts overflow or agents stall,
Follow my carrots — I note them all.
Happy debugging from this fluffy machine!

🚥 Pre-merge checks | ✅ 3 | ❌ 2

❌ Failed checks (2 warnings)

Check name Status Explanation Resolution
Linked Issues check ⚠️ Warning The PR significantly exceeds the scope requested in issue #326, which asked for a minimal docs-only addition with a one-line description and link to the WFGY Problem Map reference. Review whether the comprehensive troubleshooting guide (covering memory, agent harness, hooks, installation, performance, errors) aligns with the issue's intent for a compact, minimal reference document, or consider whether additional issue creation was needed.
Out of Scope Changes check ⚠️ Warning The PR introduces extensive troubleshooting content across multiple issue categories beyond the specific failure patterns mentioned in issue #326 (memory/context drift, stale context, semantic/operational mismatches, answer inconsistency, instruction confusion, reasoning breakdowns). Clarify whether installation, performance optimization, and common error message sections were intended, or if the scope should be narrowed to focus primarily on the agent harness workflow failures and context-related issues from issue #326.
✅ Passed checks (3 passed)
Check name Status Explanation
Description Check ✅ Passed Check skipped - CodeRabbit’s high-level summary is enabled.
Title check ✅ Passed The title clearly summarizes the main change: adding a comprehensive troubleshooting guide documentation file for the ECC plugin, which aligns with the PR's core objective.
Docstring Coverage ✅ Passed No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.

✏️ Tip: You can configure your own custom pre-merge checks in the settings.

✨ Finishing Touches
🧪 Generate unit tests (beta)
  • Create PR with unit tests
  • Post copyable unit tests in a comment

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

Comment @coderabbitai help to get the list of available commands and usage tips.

Copy link

@cubic-dev-ai cubic-dev-ai bot left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

2 issues found across 1 file

Prompt for AI agents (unresolved issues)

Check if these issues are valid — if so, understand the root cause of each and fix them. If appropriate, use sub-agents to investigate and fix each issue separately.


<file name="TROUBLESHOOTING.md">

<violation number="1" location="TROUBLESHOOTING.md:291">
P1: Comment says "Archive old observations" but the command permanently deletes them with `-delete`. This is misleading and risks unintended data loss. Either rename the comment to reflect deletion, or change the command to actually archive (e.g., move/compress the files).</violation>

<violation number="2" location="TROUBLESHOOTING.md:335">
P2: `chmod -R 755` makes all data files (e.g., `.jsonl`) executable, which is a security anti-pattern. Use `chmod -R u+rwX,go+rX` instead — the uppercase `X` only sets execute on directories, not regular files.</violation>
</file>

Reply with feedback, questions, or to request a fix. Tag @cubic-dev-ai to re-run a review.

Copy link
Contributor

@coderabbitai coderabbitai bot left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Actionable comments posted: 5

🧹 Nitpick comments (3)
TROUBLESHOOTING.md (3)

156-156: Simplify pipeline (useless use of cat).

The command can be simplified:

-cat ~/.claude/settings.json | grep -A 10 '"hooks"'
+grep -A 10 '"hooks"' ~/.claude/settings.json
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.

In `@TROUBLESHOOTING.md` at line 156, Replace the useless use of cat in the
pipeline by running grep directly on the file; change the command `cat
~/.claude/settings.json | grep -A 10 '"hooks"'` to `grep -A 10 '"hooks"'
~/.claude/settings.json` so grep reads the file directly and avoids the
unnecessary cat process.

271-272: Consider adding a warning about lock file deletion.

Deleting the lock file can lead to different dependency versions being installed. While this is intentional when switching package managers, a brief warning would be helpful.

📝 Suggested addition
 # Remove conflicting lock files
+# Warning: This may cause different versions to be installed
 rm package-lock.json  # If using pnpm/yarn/bun
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.

In `@TROUBLESHOOTING.md` around lines 271 - 272, Update the "Remove conflicting
lock files" section to include a brief warning that deleting lockfiles (the
existing "rm package-lock.json" instruction) can change installed dependency
versions; advise making a backup or committing the lockfile before removal,
mention this is typically only needed when switching package managers
(pnpm/yarn/bun), and suggest running a fresh install and checking CI/local
builds after removal to ensure reproducible results.

35-35: Simplify pipeline (useless use of cat).

The command can be simplified by removing the unnecessary cat:

-cat large-file.txt | head -50
+head -50 large-file.txt
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.

In `@TROUBLESHOOTING.md` at line 35, Replace the unnecessary pipeline "cat
large-file.txt | head -50" with a direct head invocation to simplify the
command; change the usage to call head directly (e.g., head -n 50
large-file.txt) so you avoid the useless use of cat.
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.

Inline comments:
In `@TROUBLESHOOTING.md`:
- Around line 296-297: Update the "Use local caching" section that contains the
line "Enable Redis for semantic search caching" to either (A) expand it with
concise, actionable setup steps: list installation commands for common OSes or
Docker, required Redis configuration (host/port/auth), how to set the related
environment variables used by the app (e.g., REDIS_URL or
REDIS_HOST/REDIS_PORT), and an example config snippet and verification command
(redis-cli PING), or (B) remove the Redis mention entirely if Redis support is
not implemented; ensure you update any cross-references in the doc. Target the
heading text "Use local caching" and the specific line "Enable Redis for
semantic search caching" when making the change.
- Around line 234-235: Add a clear warning and backup recommendation above the
destructive command "rm -rf ~/.claude/plugins/cache/*" in TROUBLESHOOTING.md:
state that this will permanently delete plugin cache and advise users to create
a backup (suggest commands such as copying or archiving the
~/.claude/plugins/cache directory) or to inspect contents first, and present an
alternative non-destructive step (e.g., moving the cache to a backup location)
so users can recover if needed.
- Around line 5-12: Add the two missing TOC entries for the sections titled
"Common Error Messages" and "Getting Help" by inserting bullet links "- [Common
Error Messages](`#common-error-messages`)" and "- [Getting Help](`#getting-help`)"
into the existing Table of Contents (the list under "## Table of Contents") so
the TOC includes those two sections (titles "Common Error Messages" and "Getting
Help") that appear later in the document.
- Around line 290-291: The current command 'find ~/.claude/homunculus -name
"observations.jsonl" -size +10M -delete' is destructive; replace it with an
archival workflow that moves matching files to a dated/archive directory and
optionally compresses them instead of deleting; implement a find+mkdir+mv (or
rsync) pattern that creates ~/.claude/homunculus/archive/<YYYYMMDD>/, moves each
observations.jsonl there (preserving path or renaming with a timestamp), and
then gzip or tar the archived files to reclaim space while keeping a backup for
recovery and auditing.
- Around line 57-60: Update the two commands in TROUBLESHOOTING.md that build
observation file paths using $(basename $PWD) — they should reference ECC
project hash IDs stored in ~/.claude/homunculus/projects.json; replace the
incorrect paths in the lines containing "tail -20
~/.claude/homunculus/$(basename $PWD)/observations.jsonl" and "rm
~/.claude/homunculus/$(basename $PWD)/observations.jsonl" with either the
wildcard form "…/homunculus/*/observations.jsonl" or instruct to first look up
the 12-character project hash in ~/.claude/homunculus/projects.json and then
construct the path as
"~/.claude/homunculus/projects/<project-hash>/observations.jsonl".

---

Nitpick comments:
In `@TROUBLESHOOTING.md`:
- Line 156: Replace the useless use of cat in the pipeline by running grep
directly on the file; change the command `cat ~/.claude/settings.json | grep -A
10 '"hooks"'` to `grep -A 10 '"hooks"' ~/.claude/settings.json` so grep reads
the file directly and avoids the unnecessary cat process.
- Around line 271-272: Update the "Remove conflicting lock files" section to
include a brief warning that deleting lockfiles (the existing "rm
package-lock.json" instruction) can change installed dependency versions; advise
making a backup or committing the lockfile before removal, mention this is
typically only needed when switching package managers (pnpm/yarn/bun), and
suggest running a fresh install and checking CI/local builds after removal to
ensure reproducible results.
- Line 35: Replace the unnecessary pipeline "cat large-file.txt | head -50" with
a direct head invocation to simplify the command; change the usage to call head
directly (e.g., head -n 50 large-file.txt) so you avoid the useless use of cat.

ℹ️ Review info
⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro

Run ID: 753e653d-0165-414f-8e2f-8df03f0a65b0

📥 Commits

Reviewing files that changed from the base of the PR and between af51fca and 1c6e401.

📒 Files selected for processing (1)
  • TROUBLESHOOTING.md

Copy link

Copilot AI left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Pull request overview

Adds a new troubleshooting reference for the Everything Claude Code (ECC) plugin to help users diagnose common installation, hook/workflow, memory/context, and performance issues.

Changes:

  • Introduces TROUBLESHOOTING.md with symptom/cause/solution sections for frequent ECC failure modes.
  • Adds command-line diagnostic steps for hooks, plugin installation, and observation storage.
  • Documents common error messages and escalation steps (“Getting Help”).

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Comment on lines +245 to +248
# Manual install (if marketplace fails)
git clone https://github.com/affaan-m/everything-claude-code.git
cp -r everything-claude-code ~/.claude/plugins/ecc
```
Copy link

Copilot AI Mar 10, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The “Manual install” steps here don’t match the manual installation instructions in README.md (which copy agents/commands/skills/rules into ~/.claude/{agents,commands,skills,rules} rather than copying the repo into ~/.claude/plugins/ecc). Please align this section with the documented manual install flow so users don’t end up with an unsupported directory layout.

Copilot uses AI. Check for mistakes.
Comment on lines +290 to +292
# Archive old observations
find ~/.claude/homunculus -name "observations.jsonl" -size +10M -delete

Copy link

Copilot AI Mar 10, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This claims to “Archive old observations” but the command permanently deletes any observations.jsonl over 10MB. That’s a destructive operation and can wipe both global and project-scoped histories. Prefer a safer rotation/backup approach (e.g., move to a timestamped file) and explicitly call out the data-loss risk if deletion is intended.

Copilot uses AI. Check for mistakes.
Comment on lines +311 to +312
# Check for runaway processes
top -o cpu | grep claude
Copy link

Copilot AI Mar 10, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

top -o cpu | grep claude generally won’t work as written because top is interactive and doesn’t reliably emit grep-able output on many platforms. Use a non-interactive command for this check (e.g., ps filtered by process name, or platform-specific non-interactive top flags) so the troubleshooting step is executable.

Copilot uses AI. Check for mistakes.
Comment on lines +342 to +347
cd ~/.claude/plugins/cache/everything-claude-code
npm install

# Or for manual install
cd ~/.claude/plugins/ecc
npm install
Copy link

Copilot AI Mar 10, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This repeats the ~/.claude/plugins/ecc manual-install path, but the repo’s README manual install instructions don’t place anything under ~/.claude/plugins/. Align the troubleshooting steps with the documented manual install locations so the npm install advice points at a real directory (or clarify when users would have a local plugin checkout under ~/.claude/plugins/).

Copilot uses AI. Check for mistakes.
Comment on lines +5 to +11
## Table of Contents

- [Memory & Context Issues](#memory--context-issues)
- [Agent Harness Failures](#agent-harness-failures)
- [Hook & Workflow Errors](#hook--workflow-errors)
- [Installation & Setup](#installation--setup)
- [Performance Issues](#performance-issues)
Copy link

Copilot AI Mar 10, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The Table of Contents omits sections that exist later in the document (e.g., “Common Error Messages”, “Getting Help”, “Related Documentation”), which makes navigation harder. Add these sections to the TOC (or remove the TOC) so anchors stay complete/accurate.

Copilot uses AI. Check for mistakes.
Comment on lines +53 to +60
# Check if observations are being recorded
ls ~/.claude/homunculus/*/observations.jsonl

# View recent observations
tail -20 ~/.claude/homunculus/$(basename $PWD)/observations.jsonl

# Reset observations if corrupted
rm ~/.claude/homunculus/$(basename $PWD)/observations.jsonl
Copy link

Copilot AI Mar 10, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The observation file paths here don’t match the repo’s documented continuous-learning-v2 layout (project-scoped ~/.claude/homunculus/projects/<project-hash>/observations.jsonl with a global fallback at ~/.claude/homunculus/observations.jsonl). Using ~/.claude/homunculus/*/… and $(basename $PWD) will usually point to a non-existent directory. Update the commands to reference the correct projects/<project-hash> path (and consider pointing users to ~/.claude/homunculus/projects.json to map project names to hashes).

Copilot uses AI. Check for mistakes.
Comment on lines +111 to +113
# 3. Set shorter timeouts
export CLAUDE_TIMEOUT=30

Copy link

Copilot AI Mar 10, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

CLAUDE_TIMEOUT isn’t referenced anywhere in this repository (no code/docs mention it besides this line), so readers may assume it’s a supported knob when it isn’t. Either remove this suggestion or replace it with the actual supported way to adjust timeouts (with a link to the relevant upstream Claude Code docs, if this is an external env var).

Copilot uses AI. Check for mistakes.
find ~/.claude/plugins -name "*.sh" -exec chmod +x {} \;

# Fix observation directory permissions
chmod -R 755 ~/.claude/homunculus
Copy link

Copilot AI Mar 10, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The command chmod -R 755 ~/.claude/homunculus weakens filesystem permissions on the homunculus directory, making all observation data and related files readable and traversable by other local users on multi-user systems. Since these observation logs can contain sensitive code, commands, and context, another user on the same machine could access them simply by browsing the directory. Instead, keep this directory restricted to the owning user (e.g., using owner-only permissions and avoiding recursive world/group read/execute) and adjust only the minimal paths necessary.

Copilot uses AI. Check for mistakes.
Copy link
Owner

@affaan-m affaan-m left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Automated review: doc-only changes look good. Approving.

@affaan-m affaan-m merged commit 5644415 into affaan-m:main Mar 11, 2026
38 of 39 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Suggestion: add an optional troubleshooting reference for memory, context, and agent harness workflow failures

3 participants