home / skills / yonatangross / orchestkit / feedback

feedback skill

safe

This skill helps you manage the OrchestKit feedback system, enabling status checks, pausing, exporting data, and privacy settings for better customization.

npx playbooks add skill yonatangross/orchestkit --skill feedback

Review the files below or copy the command above to add this skill to your agents.

Files (6)

SKILL.md

2.8 KB

---
name: feedback
license: MIT
compatibility: "Claude Code 2.1.34+."
description: "Manages OrchestKit feedback system. Use when providing feedback or viewing usage analytics."
argument-hint: "[subcommand]"
context: inherit
version: 1.2.0
author: OrchestKit
tags: [feedback, learning, patterns, metrics, privacy, analytics, consent]
user-invocable: true
allowed-tools: [Read, Write, Edit, Grep, Glob]
complexity: low
metadata:
  category: document-asset-creation
---

# Feedback - Manage Learning System

View and manage the OrchestKit feedback system that learns from your usage.

## Overview

- Checking feedback system status
- Pausing/resuming learning
- Resetting learned patterns
- Exporting feedback data
- Managing privacy settings
- Enabling/disabling anonymous analytics sharing
- Viewing privacy policy

## Usage

```
/feedback                    # Same as status
/feedback status             # Show current state
/feedback pause              # Pause learning
/feedback resume             # Resume learning
/feedback reset              # Clear learned patterns
/feedback export             # Export feedback data
/feedback settings           # Show/edit settings
/feedback opt-in             # Enable anonymous sharing
/feedback opt-out            # Disable anonymous sharing
/feedback privacy            # View privacy policy
/feedback export-analytics   # Export anonymous analytics for review
```

## Subcommands

| Subcommand | Description |
|------------|-------------|
| `status` (default) | Show current feedback system state, learned patterns, agent performance |
| `pause` | Temporarily pause learning without clearing data |
| `resume` | Resume paused learning |
| `reset` | Clear all learned patterns (requires "RESET" confirmation) |
| `export` | Export all feedback data to `.claude/feedback/export-{date}.json` |
| `settings` | Show/edit settings (usage: `/feedback settings <key> <value>`) |
| `opt-in` | Enable anonymous analytics sharing (GDPR-compliant consent) |
| `opt-out` | Disable anonymous analytics sharing (revokes consent) |
| `privacy` | Display the full privacy policy |
| `export-analytics` | Export anonymous analytics to file for review before sharing |

**Output:** Each subcommand displays formatted status, confirmation prompts, or exported file paths. See [Subcommand Reference](references/subcommand-reference.md) for detailed actions and expected output for each subcommand.

## Consent and Security

See [Consent and Security Rules](rules/consent-and-security.md) for GDPR consent management, security restrictions, and analytics data sharing policies.

## Related Skills

- skill-evolution: Evolve skills based on feedback

## File Locations

See [File Locations](references/file-locations.md) for storage details.

See [Privacy Policy](references/privacy-policy.md) for full privacy documentation.

Overview

This skill manages the OrchestKit feedback system that learns from agent usage and performance. It provides commands to inspect status, control learning (pause/resume/reset), export feedback data and analytics, and manage privacy and consent. Use it to audit learned patterns, export data for review, and control anonymous analytics sharing. The skill is designed for safe, GDPR-compliant feedback handling.

How this skill works

The skill exposes simple subcommands to query and control the feedback system: status, pause, resume, reset, export, settings, opt-in, opt-out, privacy, and export-analytics. Each subcommand returns human-readable status, prompts for required confirmations (for destructive actions), and writes exported files to a predictable path. Consent and privacy toggles record GDPR-compliant choices and affect what anonymous analytics may be shared.

When to use it

Audit what the feedback system has learned about agents and patterns
Pause learning during sensitive testing or deployments
Reset learned patterns when retraining or changing agent logic
Export feedback or anonymous analytics for external review or compliance
Change privacy or consent settings to comply with organizational policy

Best practices

Check status before pausing or resetting to avoid unintended data loss
Use pause when testing to retain data but stop new learning
Require explicit confirmation for destructive actions (e.g., RESET) and follow the prompt
Export analytics to a secure location and review before sharing externally
Maintain documentation of opt-in/opt-out decisions for compliance audits

Example use cases

/feedback status to review current learning state and recent patterns
/feedback pause during a live experiment to prevent learning from transient behavior
/feedback reset after major model changes to clear outdated patterns
/feedback export-analytics to generate an anonymized dataset for privacy review
/feedback settings consent true to record GDPR-compliant analytics consent

FAQ

What happens when I run /feedback reset?

Reset clears all learned patterns and performance summaries; the command requires an explicit RESET confirmation prompt to prevent accidental data loss.

Does opt-in share personal data?

No. Opt-in enables anonymized analytics sharing following GDPR-compliant rules; personal identifiers are not included in exported anonymous analytics.