SkillAgentSearch skills...

harness-doctor

Diagnose the harness itself. Reports version drift vs upstream, plugin staleness, MCP version pins, hook log size, memory growth, settings.json size. Run monthly or when behavior feels off.

Install / Use

npx skills add qyndex/claude-init

Installs into whichever agent you are using.

About this skill
⚡

Claude Commands

Claude Code slash commands

Quality Score

62/100

Supported Platforms

Claude Code

Our assessment of harness-doctor

harness-doctor scores 62/100 on our quality scale, 4042nd of 4,611 Development & Engineering skills we index.

Its Claude Commands is 2.6 KB long, well organised into 11 sections with 1 code example: a solid amount of guidance for an agent.

It has no GitHub stars yet, so there is no community track record; judge it on its content.

Substance
26/30
Structure
17/20
Description
15/15
Adoption
0/20
Freshness
5/15

Maintenance, license and trust

  • We could not determine when the repository was last updated.
  • Our last check on 2026-10-20 found the source still online.
  • No license is declared. By default that means all rights are reserved: you can read it, but reusing or redistributing it is not clearly permitted. Ask the author before building on it commercially.
  • Its trust signals score 68/100, with 3 cautions from licensing, adoption, age or documentation. These come from repository metadata, not a code audit — read the skill file before letting an agent act on it.

Safety scan

No issues found

Our scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands. An AI review of the same text found nothing harmful.

AI review by kimi-k2.7-code on 2026-09-24. Automated pattern scan on 2026-09-24. It catches known dangerous patterns, not every risk — read a skill before letting an agent act on it.

harness-doctor compared with similar skills

All 4 of these similar skills score higher than harness-doctor; compare them before choosing.

SkillScoreStarsUpdatedFormat
harness-doctor (this skill)by qyndex620—Claude Commands
claude-memby thedotmack100113.6ktodayCLAUDE.md
taste-skillby Leonxlnx100110.8k1d agoCLAUDE.md
Agent-Reachby Panniantong100109.3ktodayCLAUDE.md
Understand-Anythingby Egonex-AI10091.0k11d agoCLAUDE.md

Frequently asked questions

How do I install harness-doctor?
Run npx skills add qyndex/claude-init. The install tabs above show the steps for each supported agent.
Which AI agents does harness-doctor work with?
It is written for Claude Code, as a Claude Commands file. Other agents that read the same format can often use it too.
Is harness-doctor safe to use?
Our scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands. An AI review of the same text found nothing harmful. It declares no license and scores 68/100 on trust signals. Skills are instructions an agent will follow, so read the file before installing it and do not approve commands you do not understand.
Is harness-doctor still maintained?
We could not determine when the repository was last updated.

description: Diagnose the harness itself. Reports version drift vs upstream, plugin staleness, MCP version pins, hook log size, memory growth, settings.json size. Run monthly or when behavior feels off. argument-hint: "[--fix]" allowed-tools: Read, Glob, Grep, Bash, WebFetch disable-model-invocation: true

/harness-doctor — Self-diagnostic

ROOT="$(cd .claude/.. && pwd)"

echo "# Harness Doctor — $(date -Iseconds)"
echo

# 1. Version check
echo "## Version"
if [ -f .claude/VERSION ]; then
  cat .claude/VERSION
else
  echo "⚠ .claude/VERSION missing — cannot determine harness version"
fi

# 2. Compare to upstream (if origin set)
if [ -f .claude/VERSION ]; then
  upstream=$(grep 'source=' .claude/VERSION | cut -d= -f2)
  if [ -n "$upstream" ]; then
    echo
    echo "## Upstream drift"
    echo "Source: $upstream"
    latest=$(claude -p "WebFetch $upstream/raw/main/.claude/VERSION and return the harness_version field" --bare 2>/dev/null)
    echo "Latest: $latest"
  fi
fi

# 3. Plugin staleness
echo
echo "## Plugin staleness"
if command -v claude >/dev/null; then
  claude -p "/plugin list with last-update dates" --bare 2>/dev/null | head -30
fi

# 4. MCP version pins
echo
echo "## MCP server pins"
echo "Pinned to @latest (supply-chain risk):"
grep -h '@latest' .mcp.json 2>/dev/null | head -10
echo
echo "Pinned to specific versions (good):"
grep -h '@[0-9]' .mcp.json 2>/dev/null | head -10

# 5. Hook log size
echo
echo "## Hook log size"
if [ -d .claude/hooks/.log ]; then
  du -sh .claude/hooks/.log/* 2>/dev/null | sort -rh | head -10
fi

# 6. Memory file size
echo
echo "## Memory file sizes"
for f in .claude/memory/MEMORY.md .claude/memory/decisions/*.md .claude/memory/incidents/*.md; do
  [ -f "$f" ] && wc -l "$f" 2>/dev/null
done | sort -rn | head -20

# 7. Settings.json size
echo
echo "## Settings.json metrics"
echo "Lines: $(wc -l < .claude/settings.json)"
echo "Allow patterns: $(jq -r '.permissions.allow | length' .claude/settings.json)"
echo "Ask patterns: $(jq -r '.permissions.ask | length' .claude/settings.json)"
echo "Deny patterns: $(jq -r '.permissions.deny | length' .claude/settings.json)"

# 8. Validate
echo
echo "## Validate"
bash .claude/scripts/validate.sh 2>&1 | tail -3

# 9. Recommend
echo
echo "## Recommendations"
echo "- If MCP @latest pins → run /audit-plugins to pin"
echo "- If MEMORY.md > 200 lines → run /dream"
echo "- If settings.json > 500 lines → /audit-settings to prune unused entries"
echo "- If upstream drift → git remote update && consider pulling harness updates"

Recurring scan

Add to dream-cron so this runs daily and surfaces issues at session-start via additionalContext.

$ARGUMENTS

Related Skills

View on GitHub
GitHub Stars0
CategoryDevelopment
UpdatedNaNy ago
Forks0

Trust signals

68/100

From repository metadata: license, adoption, age and documentation. Not a code audit — see the Safety scan above for what the skill file itself contains.

2 medium1 low