Feature Flag Manager

Feature flag management for AI agents — toggle features, A/B testing, gradual rollouts

Security Scan

The skill's code, instructions, and resource access are coherent with a local feature-flag manager: it reads/writes a .featureflags/config.json and does not request credentials, external network access, or unusual privileges.

Purpose & Capability

Name/description match the actual behavior: the SKILL.md and flag.js implement creating, toggling, listing, and querying flags stored in .featureflags/config.json. No unrelated credentials, binaries, or APIs are requested.

Instruction Scope

Runtime instructions are limited to local file-based feature-flag operations and show example usage of flag.js; they do not instruct reading other system files, environment variables, or sending data externally.

Install Mechanism

No install spec is present and this is an instruction-only skill with a small JavaScript engine file; nothing is downloaded or extracted from external hosts.

Credentials

The skill declares no required environment variables, credentials, or config paths beyond writing a local .featureflags directory, which is proportional to its purpose.

Persistence & Privilege

always is false and the skill does not modify other skills or system-wide settings; it only writes its own config in the working directory (.featureflags), which is expected for this functionality.

Assessment

This skill appears to be a simple, local feature-flag utility that stores flags in a .featureflags/config.json file. Before installing or using it: (1) be aware it writes to your working directory (back up or ignore .featureflags if you don't want it in a repo), (2) it has no access control—if multiple people or services run it, consider adding safeguards (ACLs) or moving storage to a controlled service, (3) review and test parsing/edge cases if you plan to use it in production (argument parsing is simplistic), and (4) if you need remote/centralized feature flags, integrate with a vetted remote service rather than relying on this local file-based approach. Overall there are no signs of credential exfiltration, network calls, or other incoherent behavior.

Manage feature flags for AI agents — enable/disable features, A/B testing, gradual rollouts, and percentage-based releases.

Skill Metadata

• Slug: feature-flag-manager
• Version: 1.0.0
• Description: Feature flag management system for AI agents. Toggle features on/off, run A/B tests, gradual percentage rollouts, and control feature visibility without deploying new code.
• Category: automation
• Trigger Keywords: feature flag, feature toggle, A/B testing, gradual rollout, release management, canary release, percentage rollout

---

Capabilities

1. Create Feature Flag

# Create a simple on/off flag
node flag.js create my-feature --description "New dashboard"

# Create percentage rollout (initially 10%)
node flag.js create dark-mode --percentage 10 --description "Dark theme"

# Create A/B test variant
node flag.js create pricing-page --variants control,v1,v2 --weights 50,25,25

• Stores flags in .featureflags/ JSON config
• Supports percentage-based rollout (0-100%)
• Multi-variant A/B testing with custom weight distribution

2. Check Flag Status

# Check if feature is enabled
node flag.js enabled my-feature

# Get variant for current user (for A/B)
node flag.js variant pricing-page --user-id user123

# Get rollout percentage
node flag.js percentage dark-mode

• Returns boolean for simple flags
• Returns variant name for A/B tests
• Returns rollout percentage

3. Update Flag

# Enable/disable immediately
node flag.js toggle my-feature

# Update rollout percentage (gradual increase)
node flag.js update dark-mode --percentage 50

# Pause a flag
node flag.js pause pricing-page

• Instant toggle for emergency rollback
• Percentage updates for gradual rollout
• Pause preserves configuration

4. List & Monitor

# List all flags with status
node flag.js list

# Show flag history
node flag.js history my-feature

# Export flags for reporting
node flag.js export --format json

• Real-time overview of all flags
• Change history with timestamps
• Export for analytics integration

---

Configuration

// .featureflags/config.json
{
  "flags": {
    "new-dashboard": {
      "enabled": true,
      "percentage": 100,
      "variants": null,
      "description": "Redesigned dashboard UI"
    },
    "dark-mode": {
      "enabled": true,
      "percentage": 25,
      "variants": null,
      "description": "Dark theme support"
    },
    "pricing-page": {
      "enabled": true,
      "percentage": 100,
      "variants": ["control", "v1", "v2"],
      "weights": [50, 25, 25]
    }
  }
}

Use Cases

1. Gradual Rollout: Start with 5% users, increase to 100% over time
2. A/B Testing: Test different UX variants and measure conversion
3. Kill Switch: Instantly disable buggy feature without deployment
4. User Segmentation: Target specific user groups
5. Remote Configuration: Change behavior without code changes

Integration Example

// In your agent code
const flag = require('./flag.js');

async function handleRequest(req) {
  // Check if new feature is enabled
  if (await flag.enabled('new-dashboard', { userId: req.userId })) {
    return renderNewDashboard(req);
  }
  return renderLegacyDashboard(req);
}