Repository navigation
AnythingLLM AgentFlow Workflow Examples
Practical workflow examples showing how to chain AnythingLLM AgentFlows for complete security assessment sequences - from initial recon through email OSINT, breach checking, threat intel, and final report generation.
Demonstrates multi-flow chains: DomainRecon -> EmailOSINT -> BreachChecker -> ThreatIntelCheck -> VulnReportGenerator. Includes example prompts, expected output structures, and guidance on passing data between flows.
Enable end-to-end automated security assessments using chained AgentFlows with minimal manual intervention between steps.
Planning an automated recon workflow; combining multiple flows for a comprehensive assessment; understanding how flows connect and hand off data to each other.
This document provides practical workflow examples using AnythingLLM custom AgentFlows for cybersecurity assessments and penetration testing. These workflows demonstrate how to chain together multiple flows to accomplish common security tasks efficiently.
The following custom AgentFlows are referenced in these workflows:
- CompanyOSINT - Gathers public information about target companies
- DomainRecon - Performs comprehensive domain reconnaissance
- EmailOSINT - Investigates email addresses and usernames across platforms
- GitHubSecrets - Searches GitHub for exposed credentials and sensitive data
- BreachChecker - Checks domains/emails against known data breaches
- NmapAnalyzer - Analyzes Nmap scan results and provides recommendations
- CVELookup - Retrieves CVE details and exploitability information
- VulnReportGenerator - Creates professional vulnerability reports
- ThreatIntelCheck - Checks IP/domain reputation against threat intelligence
Purpose: Gather intelligence before active testing begins
Workflow Steps:
-
CompanyOSINT โ Identify company structure, employee information, and email formats
- Input:
company_name,domain
- Input:
-
DomainRecon โ Map attack surface and discover subdomains
- Input:
domain
- Input:
-
EmailOSINT โ Investigate key personnel emails for breach history
- Input: Key personnel emails from CompanyOSINT
-
GitHubSecrets โ Search for exposed credentials and secrets
- Input:
target_domain
- Input:
-
BreachChecker โ Identify compromised accounts and leaked data
- Input:
domain
- Input:
Output: Comprehensive pre-engagement intelligence report
Purpose: Complete network security assessment workflow
Workflow Steps:
-
Nmap Scan - Execute comprehensive port and service scan
nmap -sV -sC -p- target.com -oN scan.txt
-
NmapAnalyzer โ Parse scan results and prioritize findings
- Input:
scan_results(from scan.txt),target_info
- Input:
-
CVELookup โ Research each identified CVE
- Input:
cve_id,system_context - Repeat for each CVE discovered
- Input:
-
VulnReportGenerator โ Document each finding professionally
- Input:
finding_title,finding_details,affected_systems,client_name - Repeat for each vulnerability
- Input:
-
Compile Final Report - Aggregate all findings into deliverable
Output: Professional penetration test report ready for client delivery
Purpose: Rapid investigation of security incidents
Workflow Steps:
-
ThreatIntelCheck โ Assess suspicious IP addresses
- Input: Suspicious IP from logs/alerts
-
DomainRecon โ Investigate attacker domains
- Input: Attacker domain (if identified)
-
EmailOSINT โ Research associated email addresses
- Input: Email addresses from investigation
-
BreachChecker โ Identify compromised accounts
- Input: Domain or email addresses
Output: Incident analysis with threat actor attribution and impact assessment
Use Case: Fast initial security check
Nmap Scan
โ
NmapAnalyzer (prioritize findings)
โ
CVELookup (research critical CVEs)
โ
VulnReportGenerator (document findings)
Use Case: Comprehensive intelligence gathering
CompanyOSINT (company intelligence)
โ
EmailOSINT (key personnel investigation)
โ
GitHubSecrets (credential exposure)
โ
BreachChecker (breach history)
Use Case: Attack surface enumeration
DomainRecon (discover assets)
โ
Nmap Scan (active scanning)
โ
ThreatIntelCheck (verify IPs)
โ
NmapAnalyzer (analyze results)
Use Case: Batch document multiple findings
All vulnerability findings
โ
VulnReportGenerator (repeat for each)
โ
VulnReportGenerator (repeat for each)
โ
VulnReportGenerator (repeat for each)
โ
Compile into final report
Use Case: Standard reconnaissance workflow
Workflow Steps:
- CompanyOSINT โ Determine email format and company structure
- EmailOSINT โ Investigate key personnel emails for exposed data
- DomainRecon โ Map complete attack surface including subdomains
- GitHubSecrets โ Identify exposed credentials in public repositories
Output: Complete OSINT intelligence package
- Always ensure you have proper authorization before executing these workflows
- Document each step as you progress through workflows
- Use descriptive inputs for better analysis quality
- Chain workflows based on findings from previous steps
- Save outputs from each flow for final report compilation
These workflows can be:
- Modified to suit specific engagement types
- Combined for comprehensive assessments
- Abbreviated for time-sensitive tasks
- Extended with additional flows as needed
- Start broad, narrow down - Begin with OSINT, then focus on discovered assets
- Prioritize findings - Use NmapAnalyzer and CVELookup to identify critical issues first
- Document continuously - Run VulnReportGenerator for each finding immediately
- Verify results - Cross-reference flow outputs with manual testing
- Iterate as needed - Loop back through workflows when new information emerges
Scenario: External penetration test for Acme Corporation
CompanyOSINT โ company: Acme Corporation, domain: acme.com
DomainRecon โ domain: acme.com
GitHubSecrets โ target: acme.com
BreachChecker โ domain: acme.com
EmailOSINT โ emails discovered from CompanyOSINT
nmap -sV -sC -p- acme.com -oN acme-scan.txt
NmapAnalyzer โ scan_results: acme-scan.txt
ThreatIntelCheck โ IPs discovered from scan
CVELookup โ each CVE from NmapAnalyzer
Manual testing of discovered services
Additional targeted scans
VulnReportGenerator โ each finding documented
Compile all outputs into final report
Executive summary creation
Client presentation preparation
Document Version: 1.0
Last Updated: November 2025
Author: Pacific Northwest Computers
Purpose: AnythingLLM AgentFlow workflow reference guide
- README.md - AnythingLLM AgentFlow documentation index
- flow-domain_recon.md - First flow in the standard recon chain
- flow-vulnreport_gen.md - Final output flow for report generation
- ../ai_prompts.md - Prompt patterns used within each flow
๐ Source: AI/AnythingLLM/anything_llm_agentflows.md | โ๏ธ Suggest an edit | โฌ
๏ธ Back to Artificial Intelligence
- ๐บ๏ธ Browse by Topic
- ๐ All Repository Files
- ๐งญ Learning Paths
- ๐ Master Guide
- ๐ค Glossary
- โ๏ธ Legal & Disclaimer
๐ Foundations & Learning
๐ก๏ธ Defense & Investigations
- Incident Response
- Open Source Intelligence (OSINT)
- Reverse Engineering
- Operations Security (OPSEC)
- Compliance
โ๏ธ Platforms & Infrastructure
๐ก Hardware, Wireless & Field Projects
- PNWC Cyberdeck
- Hardware Testing
- Hardware Hacking
- Software-Defined Radio
- Wardriving
- uConsole
- Space Security
๐งฐ Assessment & Field Reference
๐ฆ Downloads & Project Resources
๐๏ธ Additional Topics