By saxyguy81
Quality gate plugin that ensures Claude Code completes tasks fully before stopping
A Claude Code quality gate plugin that ensures task completion before allowing the agent to stop.
RealityCheck intercepts Claude Code's Stop events and uses an AI judge to evaluate whether your request was fully completed. If not, it blocks the stop and provides specific next steps to complete the task.
# Add the marketplace (one-time)
claude /plugin install realitycheck@saxyguy81-plugins
# Or via CLI
claude plugin install realitycheck@saxyguy81-plugins --scope user
# Clone the repo
git clone https://github.com/saxyguy81/realitycheck.git
cd realitycheck
# Install dependencies and build
npm install
npm run build
# Test locally with Claude Code
claude --plugin-dir ./
# Clone and setup
git clone https://github.com/saxyguy81/realitycheck.git
cd realitycheck
npm run setup
# Copy hooks to your project
cp examples/claude-hooks-config.json YOUR_PROJECT/.claude/settings.local.json
Create .claude/realitycheck.config.json in your project to customize:
{
"judge": {
"model": "opus",
"timeout": 30000
},
"limits": {
"maxConsecutiveFailures": 3,
"maxTotalAttempts": 5,
"noProgressThreshold": 3
},
"debug": false
}
/clear or session resumeclaude plugin uninstall realitycheck
cd /path/to/realitycheck
npm run uninstall
Remove .claude/settings.local.json from your projects.
# Run tests
npm test
# Run with coverage
npm run test:coverage
# Type check
npm run typecheck
# Lint
npm run lint
# Full check
npm run check
MIT
Executes bash commands
Hook triggers when Bash tool is used
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
npx claudepluginhub saxyguy81/realitycheckReliability hooks for Claude Code - prevents common mistakes and enforces quality standards
Automatically evaluates whether Claude should continue working instead of stopping prematurely using Claude-judged decision making
An easy way to force agent to think again if it's statement that the "Job is done and production ready" is actually done - usually it's not. Thanks to this command you don't have to check after the agent if they did their job.
Auto-runs a critical session post-mortem when Claude Code stops. Ships the analyze-session command, session-analyzer agent, and a Stop hook that triggers the review in-session (no external API calls).
Claude Code hooks for enforcing best practices and workflow automation
Enforce mandatory pre-action verification checkpoints to prevent pattern-matching from overriding explicit reasoning. Use this skill when about to execute implementation actions (Bash, Write, Edit) to verify hypothesis-action alignment. Blocks execution when hypothesis unverified or action targets different system than hypothesis identified. Critical for preventing cognitive dissonance where correct diagnosis leads to wrong implementation.