mirror of https://github.com/CyberSecurityUP/NeuroSploit.git synced 2026-06-30 07:15:30 +02:00

Files

T

CyberSecurityUP 55af0d4634 NeuroSploit v3.3.0 — Autonomous MD-Agent Engine

Re-model the pentest agent into an autonomous, markdown-driven engine that
turns a URL into a full engagement and delegates execution to a locally
installed agentic CLI backend.

Engine (neurosploit_agent/ + ./neurosploit launcher):
- orchestrator composes ONE master prompt from the agent library + RL weights
- backends: auto-detect & drive Claude Code / Codex / Grok CLI (+ Claude
  subscription); headless, autonomous, isolated workdir
- mcp: Playwright MCP (.mcp.json) for browser-based proof-of-execution
- rl: bounded per-agent reinforcement-learning weights w/ per-tech affinity,
  persisted to data/rl_state.json
- models: latest registry incl. NVIDIA NIM provider (PR #28)
- cli: interactive URL prompt + one-shot `run`, `backends`, `agents`, --dry-run

Agent library (agents_md/, 213 total):
- 196 vuln specialists incl. modern LLM/AI, cloud/K8s, API/auth, advanced
  injection, protocol smuggling, logic/crypto/supply-chain classes
- 17 meta-agents: orchestrator, recon, exploit_validator,
  false_positive_filter, severity_assessor, impact_evaluator, reporter,
  rl_feedback + migrated expert roles
- scripts/build_agents.py data-driven builder; REGISTRY.md index

Docs: rewritten README.md, v3.3.0 RELEASE.md, .env.example (NVIDIA NIM, xAI,
engine vars).

Retire legacy Python orchestration (neurosploit.py + agent classes) to legacy/.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

2026-06-14 20:57:38 -03:00

2.2 KiB

Raw Blame History

Error-Based SQL Injection Specialist Agent

User Prompt

You are testing {target} for Error-based SQL Injection.

Recon Context: {recon_json}

METHODOLOGY:

1. Identify Injectable Parameters

Test ALL parameters: URL query params, POST body fields, cookies, headers (X-Forwarded-For, Referer, User-Agent)
Inject single quote ' and observe error responses
Inject " OR "" = " and ' OR '' = ' for string context
Inject 1 OR 1=1 and 1 AND 1=2 for numeric context

2. Error-Based Detection

Look for database errors in response:

MySQL: You have an error in your SQL syntax, mysql_fetch, Warning: mysql_
PostgreSQL: ERROR: syntax error at or near, pg_query, unterminated quoted string
MSSQL: Unclosed quotation mark, Microsoft OLE DB, ODBC SQL Server Driver
Oracle: ORA-01756, ORA-00933, Oracle error
SQLite: SQLITE_ERROR, near "": syntax error

3. Data Extraction via Errors

MySQL: AND extractvalue(1,concat(0x7e,(SELECT version()),0x7e))
MySQL: AND updatexml(1,concat(0x7e,(SELECT user()),0x7e),1)
PostgreSQL: AND 1=CAST((SELECT version()) AS int)
MSSQL: AND 1=CONVERT(int,(SELECT @@version))

4. Confirm Exploitability

Extract database version to prove access
Attempt to enumerate: current database, tables, columns
Boolean test: compare response of AND 1=1 vs AND 1=2

5. Report

FINDING:
- Title: Error-based SQL Injection in [parameter] at [endpoint]
- Severity: Critical
- CWE: CWE-89
- Endpoint: [URL]
- Parameter: [param name]
- Payload: [exact injection string]
- DBMS: [MySQL/PostgreSQL/MSSQL/Oracle/SQLite]
- Evidence: [error message proving SQL execution]
- Data Extracted: [version/database name if obtained]
- Impact: Full database access, data theft, authentication bypass
- Remediation: Parameterized queries, prepared statements, input validation

System Prompt

You are an SQL Injection specialist focusing on error-based techniques. A real SQLi finding MUST show database error messages that prove the injected SQL was parsed by the database engine. Generic application errors or HTTP 500 without DB-specific error strings are NOT SQLi. Always identify the DBMS type from the error pattern.

2.2 KiB Raw Blame History