AI Key Configuration
Configure your Claude API key to unlock full agentic orchestration, AI-written reports, and context-aware payload generation.
Prerequisites
- Anthropic account at console.anthropic.com
- API key starting with
sk-ant- - Sufficient API credits (Claude Sonnet usage ~$3–15 per full pentest engagement)
- Internet access from the PhantomYerra machine (not required in air-gapped mode)
-
1
Get Your Claude API Key
Log into console.anthropic.com. Navigate to API Keys → Create Key. Name it "PhantomYerra". Copy the key — it starts with
sk-ant-api03-….Warning: You only see the full key once. Copy it immediately and store it securely. If lost, create a new key. -
2
Enter Key in PhantomYerra
Go to Settings → AI Configuration → Anthropic Claude. Paste your API key and click Save & Validate.
💡 The key is immediately encrypted with AES-256-GCM and stored atdata/ai_keys.enc. It is never stored in plaintext or logged. -
3
Validation Test
PhantomYerra sends a minimal test request to the Claude API to confirm the key is valid and has credits. You'll see:
✅ Claude API key validated Model : claude-sonnet-4-20250514 Status : Active Credits: Sufficient -
4
Confirm AI Features Are Unlocked
On the Home Screen, the Automated AI Mode card should now show "AI Active" with a purple indicator. The AI orchestration toggle is available in the Mission Control Wizard.
What You Unlock with a Claude Key
- Full Automated AI Pentesting mode — Claude plans and executes the entire engagement
- Context-aware payload generation for XSS, SQLi, SSRF, and more
- Professional PoC narrative writing with business impact analysis
- Business logic vulnerability discovery via AI interview + testing
- Attack chain correlation — Claude links findings into attack paths
- Auto-remediation code generation (Terraform, Kubernetes, Python patches)
- AI-written executive reports with risk narrative
- Agentic scan orchestration — plans, adapts, and re-tests based on findings
Common Issues
Ensure you copied the full key including the sk-ant-api03- prefix. Check for leading/trailing spaces. Generate a new key from console.anthropic.com if the issue persists. Keys are invalidated if you regenerate them in the Anthropic console.
PhantomYerra detects credit exhaustion and shows a non-blocking notification. The current scan pauses the AI orchestration and continues with tool-based scanning (no AI features). Update your API key or add credits, then resume from Settings → AI Configuration → Resume Scan.
Run ollama list to see installed models. If empty, run ollama pull deepseek-r1:7b. Ensure Ollama service is running: ollama serve (or it starts automatically in most installations).
At data/ai_keys.enc — AES-256-GCM encrypted using a machine-specific key derived from the hardware fingerprint. The encryption key never leaves the machine. If you reinstall PhantomYerra on the same machine, the key can be decrypted. On a new machine, re-enter the API key.