ONE PUBLIC INTEGRATION · 24-HOUR DELIVERY

Find the dangerous assumptions before an agent does.

Get an independent, evidence-backed review of one public SKILL.md, MCP manifest, agent card, or small integration surface.

0.12SOL · ONE TIME

Trust-boundary and credential-handling review
Prompt-injection and unsafe-action findings
Payment, callback, and external-request checks where applicable
Prioritized Markdown report with concrete remediations

Pay 0.12 SOL ↗

Defensive review only. No exploitation, credential use, private-system access, or guarantee that every vulnerability will be found.

SECURITY-REVIEW.mdDELIVERABLE

01 · TRUST BOUNDARIES
02 · INPUT & PROMPT HANDLING
03 · SECRET EXPOSURE
04 · EXTERNAL REQUESTS
05 · ACTION AUTHORIZATION
06 · PAYMENT ASSUMPTIONS
07 · CALLBACK VALIDATION
08 · DATA RETENTION
09 · FAILURE MODES
10 · PRIORITY FIXES

+ EVIDENCE FOR EACH FINDING
+ CONCRETE REMEDIATIONS

AFTER PAYMENT

Send one public URL and the intended runtime.

Email the public transaction signature, the public artifact URL, intended agent/runtime, and your main concern. The 24-hour delivery window begins once those details arrive.

Open the order email →

FIXED SCOPE

Actionable findings, not a scanner dump.

Every reported issue includes evidence, impact, priority, and a practical next change.

INCLUDED

One public artifact
Up to ten findings
Priority order
Markdown delivery

NOT INCLUDED

Private-system testing
Credential use
Exploit development
Compliance certification