[RELEASE] SQLMAP SKYNET — MCP Agent Mode + RAG Memory + Live Dashboard (operator-friendly sqlmap)
Hey everyone
I’ve been building a project called SQLMAP SKYNET.
It’s still sqlmap at the core — but SKYNET adds a real operator workflow so scans are cleaner, repeatable, and easier to manage.
Only for authorized testing (your own systems / written permission).
Why it’s powerful (what makes it different)
1) MCP Agent Mode (tool-driven automation)
Instead of copy/pasting commands all day, SKYNET exposes clean actions like tools:
2) RAG Memory (remembers what actually worked)
This isn’t “AI guessing” — it’s practical memory built from your own runs:
3) Phased Scan Engine (less chaos, more control)
Scanning is broken into phases so you can stop early or go deeper only when confirmed:
4) Live Dashboard (real visibility)
Real-time logs + status so you can see exactly where it succeeds or fails.
Also supports headless usage on a server if you don’t want UI.
5) Evidence-first reporting
Outputs in multiple formats (HTML/TXT/JSON) so it’s easy to share findings and keep an audit trail.
What I’m improving next
If anyone wants screenshots or wants to test it, reply here and I’ll share details (without spamming links in the main post).
Hey everyone
I’ve been building a project called SQLMAP SKYNET.
It’s still sqlmap at the core — but SKYNET adds a real operator workflow so scans are cleaner, repeatable, and easier to manage.
Only for authorized testing (your own systems / written permission).
Why it’s powerful (what makes it different)
1) MCP Agent Mode (tool-driven automation)
Instead of copy/pasting commands all day, SKYNET exposes clean actions like tools:
- Start scan
- Run phases (Detect / Verify / Enumerate / Export)
- Stream logs
- Export reports
2) RAG Memory (remembers what actually worked)
This isn’t “AI guessing” — it’s practical memory built from your own runs:
- Stores what worked per target/environment (headers/cookies/options)
- Remembers common failure patterns (WAF blocks, rate-limit, auth missing)
- Reuses known-good profiles so you don’t restart from zero every time
3) Phased Scan Engine (less chaos, more control)
Scanning is broken into phases so you can stop early or go deeper only when confirmed:
- Detect → confirm the injection path
- Verify → confidence / stability checks
- Enumerate → DB/tables/users when allowed
- Export → evidence + reports
4) Live Dashboard (real visibility)
Real-time logs + status so you can see exactly where it succeeds or fails.
Also supports headless usage on a server if you don’t want UI.
5) Evidence-first reporting
Outputs in multiple formats (HTML/TXT/JSON) so it’s easy to share findings and keep an audit trail.
What I’m improving next
- Multi-target input with per-target options (headers/cookies/extradata)
- Cleaner report timeline (“scan story”)
- Preset profiles (fast / safe / deep) + smarter stop conditions
If anyone wants screenshots or wants to test it, reply here and I’ll share details (without spamming links in the main post).
