// ghost brief™

Ghost briefs the AI.
The AI fixes the code.

After every scan, Ghost Brief™ converts your findings into a validated, blast-radius-aware prompt pack your AI coding agent can execute immediately. No more translating findings into instructions. No more guessing what the AI might break.

See Plans →
// how it works
Three steps. One governed pipeline.
Ghost Brief™ fits into your existing scan workflow. Run a scan. Run ghost --brief. Hand the JSON to your AI agent.
// step 01

Run any Ghost scan

Ghost POI, Fix Forecast, Conflict, or any other mode scans your codebase and produces structured findings -- severity-rated, file-attributed, blast-scored.

// step 02

Run ghost --brief

Ghost Brief™ reads your findings and converts them into a validated, blast-radius-aware prompt pack -- ghost-brief.json. Each prompt is surgical. Each prompt knows what not to touch.

// step 03

Feed it to your AI agent

Drop ghost-brief.json into Claude Code, Cursor, or Copilot. Your AI agent executes the prompts in blast-radius order -- smallest blast first. Ghost Portal™ tracks every Brief you generate.

// ghost-brief.json
What's in a Brief?
Every Brief is a structured JSON file your AI agent can read directly. Ghost sorts prompts by blast radius so your agent always starts with the safest, most surgical fixes.
  • prompts[] Sorted ascending by blast score -- surgical fixes first, broad changes last. Your agent works in the safest order possible.
  • blast_radius surgical / moderate / broad -- how many parts of the codebase this change could affect.
  • files.primary The exact files to modify. Your agent reads this list and stays inside it.
  • files.do_not_touch Globs your agent must never modify -- vendor code, core framework files, third-party modules.
  • validation_hints Checklist items your agent can verify after applying the fix -- no existing tests break, change is scoped to listed files.
  • severity critical / high / medium / low -- inherited from the original scan finding.
  • source_mode Which Ghost scan produced this finding -- poi, conflict, fix-forecast, and more.
  • estimated_agent_hours Summary-level estimate across the full Brief. 1 prompt ≈ 15 minutes of AI agent time. 
// ghost-brief.json — single prompt entry
{
  "id": "GB-001",
  "title": "Missing Dependency Version Constraints",
  "severity": "low",
  "blast_radius": "moderate",
  "blast_score": 28,
  "source_mode": "poi",
  "files": {
    "primary": ["README.md"],
    "related": [],
    "do_not_touch": [
      "vendor/**",
      "app/code/Core/**"
    ]
  },
  "prompt": "Address the following issue: Missing
  Dependency Version Constraints.
  Primary file(s): README.md. Apply the
  minimal surgical fix. Do not modify
  vendor code or core framework files.",
  "validation_hints": [
    "No new errors introduced",
    "Existing tests still pass",
    "Change limited to files.primary"
  ],
  "tags": []
}
// pricing
Ghost Brief™ is included in Pro Max and above.
Upgrade once. Every scan you run automatically generates a Brief. Ghost Portal™ stores your Brief history.
Tier Price Ghost Brief™
Ghost Open Free --
Ghost Pro $99/mo --
Ghost Pro Max $199/mo ✅ Individual Brief
Ghost Team $399/mo ✅ Brief per seat
Ghost Team Max $799/mo ✅ Unified Brief
Ghost Enterprise $1,200/mo ✅ Unlimited
Ghost Enterprise Max $2,500/mo ✅ + Ghost Partner
Start with Ghost Pro Max → Compare all plans
// frequently asked
Does Ghost Brief™ write code?
No. Ghost Brief™ produces a prompt pack. Your AI coding agent executes the prompts. Ghost never touches your code -- your source files stay on your machine, and the Brief is a plain JSON file you control completely.
Which AI agents does Ghost Brief™ work with?
Any AI coding agent that accepts a prompt -- Claude Code, Cursor, GitHub Copilot, or any other. ghost-brief.json is a plain JSON file. Read it, pass the prompts to your agent, verify the hints. No proprietary integration required.
What is blast radius and why does it matter?
Blast radius measures how many parts of your codebase a change could affect. Ghost Brief™ sorts prompts by blast radius -- surgical fixes first, broad changes last -- so your AI agent works in the safest order possible. A surgical fix touches one file and poses almost no risk. A broad change ripples across the system and demands more care. Ghost handles the ordering so you don't have to think about it.