think.agentutil.net Intent Security Pre-Flight Check

{
  "action": "verify_agent_safety",
  "domain": "api",
  "context": {
    "is_test": true,
    "risk_level": "medium",
    "environment": "staging",
    "requires_approval": false
  },
  "description": "Generate a safety checklist for a test agent action to verify the pre-flight safety check system is functioning correctly before production deployment"
}

{
  "data": {
    "summary": "This is a high-risk deployment action. 2 critical, 6 high-severity, 10 total checks identified. Recommend human approval before proceeding.",
    "checklist": [
      {
        "why": "Code that works in development frequently fails in production due to environment differences, data volumes, or configuration.",
        "check": "Has this been tested in a staging/pre-production environment that mirrors production?",
        "category": "testing",
        "severity": "critical"
      },
      {
        "why": "Every production deployment must have a rollback plan. The question is not IF you'll need it but WHEN.",
        "check": "Is there a tested rollback plan? Can you revert to the previous version within minutes?",
        "category": "rollback",
        "severity": "critical"
      },
      {
        "why": "Deploying code that requires a database migration that hasn't run yet will cause immediate production errors.",
        "check": "Are all required database migrations, configuration changes, and service dependencies already in place?",
        "category": "dependencies",
        "severity": "high"
      },
      {
        "why": "Deploying during peak hours means maximum users affected if something goes wrong. Friday afternoon deploys are notorious.",
        "check": "Is this during an approved deployment window? Not during peak traffic, holidays, or critical business periods?",
        "category": "timing",
        "severity": "high"
      },
      {
        "why": "Without defined smoke tests, problems can go undetected for hours until users report them.",
        "check": "Have smoke tests been defined? What will you check immediately after deployment to verify success?",
        "category": "testing",
        "severity": "high"
      },
      {
        "why": "Support teams need to know about deployments to correlate user reports with changes.",
        "check": "Have all stakeholders (ops, support, product) been informed of the deployment?",
        "category": "notification",
        "severity": "high"
      },
      {
        "why": "A known-good snapshot enables fast rollback if the deployment introduces subtle issues found later.",
        "check": "Has the current production state been documented or snapshotted?",
        "category": "backup",
        "severity": "high"
      },
      {
        "why": "Production deployments should require explicit approval to prevent unauthorised or accidental releases.",
        "check": "Has the deployment been approved by the appropriate person (tech lead, release manager)?",
        "category": "approval",
        "severity": "high"
      },
      {
        "why": "Canary deployments limit blast radius. Deploying to 100% of users simultaneously maximises risk.",
        "check": "Is this a full deployment or a canary/gradual rollout? Should it go to a subset of users first?",
        "category": "scope",
        "severity": "medium"
      },
      {
        "why": "Feature flags allow disabling problematic features without a full rollback, reducing recovery time.",
        "check": "Are feature flags in place for new functionality that may need to be disabled without redeploying?",
        "category": "dependencies",
        "severity": "medium"
      }
    ],
    "risk_level": "high",
    "action_type": "deploy_production",
    "risk_factors": [
      "production"
    ]
  },
  "service": "https://think.agentutil.net",
  "request_id": "35ba4e16-d8b7-4895-819c-9fe8aaabbf06",
  "related_services": [
    {
      "url": "https://verify.agentutil.net",
      "name": "verify",
      "description": "Fact verification against live data sources -- exchange rates, crypto, stocks, entities, geography"
    },
    {
      "url": "https://docs.agentutil.net",
      "name": "docs",
      "description": "Search pre-indexed platform documentation -- semantic vector search across 10+ platforms"
    },
    {
      "url": "https://convert.agentutil.net",
      "name": "convert",
      "description": "Unit conversion (80+ units), live currency rates, and encoding (base64, hex, URL)"
    },
    {
      "url": "https://weather.agentutil.net",
      "name": "weather",
      "description": "Current conditions and multi-day forecasts for any location worldwide via Open-Meteo"
    },
    {
      "url": "https://validate.agentutil.net",
      "name": "validate",
      "description": "Validate emails (RFC 5322), URLs, phone numbers (E.164), credit cards (Luhn), and JSON schemas"
    },
    {
      "url": "https://time.agentutil.net",
      "name": "time",
      "description": "Timezone conversion, current time in any timezone, date arithmetic, and epoch conversions"
    },
    {
      "url": "https://hash.agentutil.net",
      "name": "hash",
      "description": "Hash with MD5/SHA-256/SHA-512, encode/decode base64/hex/URL, generate UUIDs, identify hash algorithms"
    },
    {
      "url": "https://text.agentutil.net",
      "name": "text",
      "description": "Regex match/replace/extract, text diff, slugify, word count, and markdown/HTML/JSON transforms"
    },
    {
      "url": "https://math.agentutil.net",
      "name": "math",
      "description": "Safe expression evaluation, statistics (mean/median/stddev/percentiles), percentages, and rounding"
    },
    {
      "url": "https://geocode.agentutil.net",
      "name": "geocode",
      "description": "Forward geocoding (address to lat/lon), reverse geocoding (lat/lon to address), and haversine distance"
    },
    {
      "url": "https://dns.agentutil.net",
      "name": "dns",
      "description": "DNS record lookups (A/MX/TXT/NS), reverse DNS, WHOIS/RDAP domain data, and IP geolocation"
    },
    {
      "url": "https://undo.agentutil.net",
      "name": "undo",
      "description": "Reversibility intelligence -- check if actions can be undone, find safer alternatives, batch safety warnings"
    },
    {
      "url": "https://norm.agentutil.net",
      "name": "norm",
      "description": "Baseline reality checks -- statistical anomaly detection for values against published industry benchmarks"
    },
    {
      "url": "https://context.agentutil.net",
      "name": "context",
      "description": "Situational awareness -- timezone, holiday, business calendar, and platform status checks before taking action"
    }
  ]
}