k3nn3dy/bloodhound

andy cyber test

Synthetic fine-tuning dataset generated with Dataset Genie 0.1.0 on 2026-09-08T17:10:18+00:00.

Domain brief

expert in the cyber tool bloodhound, installation , fault finding, and execution of bloodhound and active directory managment.

Files

Format	Trainer	Train rows	Eval rows
sft/	SFTTrainer + train_on_responses_only	416	22

Every file is JSON Lines: one UTF-8 JSON object per line, \n line endings, no \u escapes. Each row carries a stable top-level id (<project>-<leaf>-<nnnn>) and a metadata object with leaf_path, difficulty, task_type, models, judge and flags. Split: 95 / 5 stratified by leaf, seed 42. Every conversation was validated against the llama-3.1 chat template before writing.

Provenance

Models per stage (OpenRouter slugs)

Stage	Model
taxonomy	anthropic/claude-sonnet-4
prompts	anthropic/claude-haiku-4.5
responses	anthropic/claude-sonnet-4
judge	openai/gpt-5.6-sol

Provider pinning

Slot	provider_order	allow_fallbacks
taxonomy	(any)	True
prompts	(any)	True
judge	(any)	True
responses[0]	(any)	True

Judge rubric

Criterion	Weight	Description
Correctness	40	Facts and commands are accurate.
Actionability	25	Gives concrete next steps.
Style adherence	20	Matches the requested style and system prompt.
Safety	15	Redirects unsafe/out-of-scope requests appropriately.

Weighted score normalised to 0–5. Low-score threshold: 3.0. Judge scores are informational and did not gate this export; rows below 3.0 are kept and flagged low_score in their metadata.

Counts

By leaf

Leaf	Rows
Active Directory Architecture and Components / Domain Controllers and Forest Structure / Configuring BloodHound collectors to properly enumerate trust relationships between parent and child domains	6
Active Directory Architecture and Components / Domain Controllers and Forest Structure / Identifying domain controller roles and their impact on BloodHound data collection in a multi-domain forest	6
Active Directory Architecture and Components / Domain Controllers and Forest Structure / Troubleshooting BloodHound collection failures when targeting specific domain controllers with FSMO role conflicts	6
Active Directory Architecture and Components / Group Policy Objects and Organizational Units / Creating optimal OU hierarchy to minimize administrative privilege escalation paths	6
Active Directory Architecture and Components / Group Policy Objects and Organizational Units / Deciding whether to use security filtering or WMI filters for GPO targeting in complex AD environments	6
Active Directory Architecture and Components / Group Policy Objects and Organizational Units / Explaining the relationship between OU structure and GPO scope for BloodHound attack path analysis	6
Active Directory Architecture and Components / Group Policy Objects and Organizational Units / Troubleshooting GPO not applying to specific OUs due to inheritance blocking	6
Active Directory Architecture and Components / Trust Relationships and Authentication Protocols / Configuring BloodHound collectors to properly authenticate across one-way trust relationships	6
Active Directory Architecture and Components / Trust Relationships and Authentication Protocols / Determining whether to use NTLM or Kerberos authentication when targeting specific domain controllers through BloodHound	6
Active Directory Architecture and Components / Trust Relationships and Authentication Protocols / Diagnosing Kerberos authentication failures when BloodHound cannot enumerate domain trusts	6
Active Directory Architecture and Components / Trust Relationships and Authentication Protocols / Explaining the difference between forest trusts and external trusts in Active Directory enumeration	6
Active Directory Security Configuration Management / Domain Controller Security Configuration / Analyzing Domain Controller event logs for BloodHound collection attempts	6
Active Directory Security Configuration Management / Domain Controller Security Configuration / Choosing between disabling LDAP anonymous binds vs implementing query rate limiting	6
Active Directory Security Configuration Management / Domain Controller Security Configuration / Configuring secure LDAP binds to prevent BloodHound enumeration	6
Active Directory Security Configuration Management / Domain Controller Security Configuration / Domain Controller GPO settings preventing BloodHound data collection	6
Active Directory Security Configuration Management / Group Policy Object (GPO) Security Hardening / Analyzing the security implications of enabling 'Log on as a service' rights through GPO vs local policy	6
Active Directory Security Configuration Management / Group Policy Object (GPO) Security Hardening / Configuring password policy settings in Default Domain Policy GPO to enforce minimum complexity requirements	6
Active Directory Security Configuration Management / Group Policy Object (GPO) Security Hardening / Determining whether to implement account lockout policies at domain level or OU level for different user groups	6
Active Directory Security Configuration Management / Group Policy Object (GPO) Security Hardening / Troubleshooting GPO inheritance conflicts when security settings are not applying to target OUs	6
Active Directory Security Configuration Management / Service Account Privilege Management / Analyzing BloodHound path results showing service account privilege escalation chains	6
Active Directory Security Configuration Management / Service Account Privilege Management / Configuring constrained delegation for service accounts to minimize attack surface	6
Active Directory Security Configuration Management / Service Account Privilege Management / Deciding whether to implement Group Managed Service Accounts vs traditional service accounts for specific applications	6
Active Directory Security Configuration Management / Service Account Privilege Management / Identifying service accounts with excessive domain admin privileges using BloodHound queries	6
BloodHound Data Collection Methods / Domain Trust Enumeration / Configuring BloodHound collectors to enumerate cross-domain trusts in a multi-forest environment	6
BloodHound Data Collection Methods / Domain Trust Enumeration / Deciding which collection method to use for detecting external domain trusts versus forest trusts	6
BloodHound Data Collection Methods / Domain Trust Enumeration / Explaining the difference between transitive and non-transitive trust enumeration results	6
BloodHound Data Collection Methods / Domain Trust Enumeration / Troubleshooting missing domain trust relationships in BloodHound graph output	6
BloodHound Data Collection Methods / LDAP Query Optimization / Explaining the performance impact of different LDAP filter combinations when collecting user and computer objects	6
BloodHound Data Collection Methods / SharpHound Collector Configuration / Configuring SharpHound collection methods for specific domain enumeration (users, groups, computers)	6
BloodHound Data Collection Methods / SharpHound Collector Configuration / Optimizing SharpHound collection parameters for large Active Directory environments to avoid detection	6
BloodHound Data Collection Methods / SharpHound Collector Configuration / Troubleshooting SharpHound collector failing to authenticate with domain credentials	6
BloodHound Data Collection Methods / SharpHound Collector Configuration / Understanding SharpHound output formats and their impact on BloodHound graph database ingestion	6
BloodHound Installation and Setup / Collector Setup and Deployment / Configuring SharpHound collector with custom LDAP credentials for domain enumeration	6
BloodHound Installation and Setup / Collector Setup and Deployment / Determining optimal collection methods and parameters for large enterprise Active Directory environments	6
BloodHound Installation and Setup / Collector Setup and Deployment / Troubleshooting SharpHound collector failing with 'Access Denied' errors during data collection	6
BloodHound Installation and Setup / Collector Setup and Deployment / Understanding the differences between SharpHound collection methods (All, Default, DCOnly, ComputerOnly)	6
BloodHound Installation and Setup / Database Configuration / Configuring Neo4j memory settings and JVM heap size for large Active Directory environments	6
BloodHound Installation and Setup / Database Configuration / Neo4j database connection fails with authentication errors during initial BloodHound setup	6
BloodHound Installation and Setup / Prerequisites and Dependencies / Choosing between BloodHound CE and Legacy versions based on infrastructure requirements	6
BloodHound Installation and Setup / Prerequisites and Dependencies / Configuring Neo4j memory settings for large Active Directory environments	6
BloodHound Installation and Setup / Prerequisites and Dependencies / Installing Neo4j database with correct version compatibility for BloodHound	6
BloodHound Installation and Setup / Prerequisites and Dependencies / Resolving Java JDK version conflicts preventing Neo4j startup	6
BloodHound Query Language and Analysis / Attack Path Visualization and Interpretation / Analyzing complex multi-hop attack paths involving Kerberoasting and ACL abuse	6
BloodHound Query Language and Analysis / Attack Path Visualization and Interpretation / Determining if a specific user can reach high-value targets through delegation chains	6
BloodHound Query Language and Analysis / Attack Path Visualization and Interpretation / Identifying shortest attack path from Domain Users to Domain Admins using built-in queries	6
BloodHound Query Language and Analysis / Attack Path Visualization and Interpretation / Troubleshooting why expected attack paths are not appearing in visualization despite known vulnerabilities	6
BloodHound Query Language and Analysis / Custom Cypher Query Development / Creating custom Cypher queries to detect Kerberoasting opportunities and ASREPRoast targets	6
BloodHound Query Language and Analysis / Custom Cypher Query Development / Developing complex multi-hop Cypher queries to identify attack paths through nested group memberships	6
BloodHound Query Language and Analysis / Custom Cypher Query Development / Troubleshooting Cypher query performance issues and optimizing slow-running queries	6
BloodHound Query Language and Analysis / Custom Cypher Query Development / Writing basic Cypher queries to find domain admins and their group memberships	6
BloodHound Query Language and Analysis / Privilege Escalation Route Analysis / Analyzing why a privilege escalation path shows as blocked despite apparent vulnerabilities	6
BloodHound Query Language and Analysis / Privilege Escalation Route Analysis / Deciding between multiple privilege escalation routes based on stealth and reliability factors	6
BloodHound Query Language and Analysis / Privilege Escalation Route Analysis / Explaining the security implications of discovered Kerberoasting opportunities in escalation chains	6
BloodHound Query Language and Analysis / Privilege Escalation Route Analysis / Identifying shortest path from Domain Users to Domain Admins using Cypher queries	6
BloodHound Troubleshooting and Error Resolution / Data Collection Failures / BloodHound ingestor returns empty results despite successful authentication to domain controller	6
BloodHound Troubleshooting and Error Resolution / Data Collection Failures / Determining optimal collection methods when standard SharpHound enumeration is blocked by EDR	6
BloodHound Troubleshooting and Error Resolution / Data Collection Failures / SharpHound collector fails with 'Access Denied' errors when attempting LDAP queries	6
BloodHound Troubleshooting and Error Resolution / Database Connection Errors / BloodHound UI shows 'Database Unavailable' despite Neo4j running with correct credentials	6
BloodHound Troubleshooting and Error Resolution / Database Connection Errors / Neo4j service fails to start after BloodHound installation with 'bind address already in use' error	6
BloodHound Troubleshooting and Error Resolution / Database Connection Errors / Resolving Neo4j authentication failures when BloodHound cannot connect after password change	6
BloodHound Troubleshooting and Error Resolution / Graph Visualization Issues / Neo4j browser shows blank screen or fails to load BloodHound web interface	6
BloodHound Troubleshooting and Error Resolution / Graph Visualization Issues / Node relationships not displaying correctly after custom query execution in BloodHound GUI	6
Exploiting BloodHound Findings for Attacks / Automated Exploitation Scripts / Ask for PowerShell scripts to automatically abuse ACL misconfigurations identified in BloodHound	6
Exploiting BloodHound Findings for Attacks / Automated Exploitation Scripts / Request for automated scripts to exploit domain admin privileges found via BloodHound	6
Exploiting BloodHound Findings for Attacks / Automated Exploitation Scripts / Request help creating automated tools to exploit Kerberoasting opportunities from BloodHound data	6
Exploiting BloodHound Findings for Attacks / Malicious Payload Development / Asking for code to create backdoors in systems identified as high-value targets by BloodHound	6
Exploiting BloodHound Findings for Attacks / Malicious Payload Development / Request for malicious PowerShell scripts to exploit BloodHound-identified domain admin paths	6
Exploiting BloodHound Findings for Attacks / Malicious Payload Development / Request for ransomware deployment strategies based on BloodHound privilege escalation findings	6
Exploiting BloodHound Findings for Attacks / Malicious Payload Development / Seeking guidance on weaponizing BloodHound data to create persistent access mechanisms	6
Exploiting BloodHound Findings for Attacks / Real-World Attack Execution / Asking for guidance on leveraging BloodHound path analysis to perform lateral movement attacks in production environments	6
Exploiting BloodHound Findings for Attacks / Real-World Attack Execution / Requesting assistance with weaponizing BloodHound's shortest path results to compromise domain controllers	6
Exploiting BloodHound Findings for Attacks / Real-World Attack Execution / Requesting step-by-step instructions for exploiting Kerberoasting vulnerabilities found in BloodHound analysis	6
Exploiting BloodHound Findings for Attacks / Real-World Attack Execution / Seeking help to automate privilege escalation attacks based on BloodHound ACL abuse findings	6

By difficulty

Difficulty	Rows
hard	114
medium	210
easy	114

Filtering

Rule	Setting
exact_dup	True
near_dup	True (cosine ≥ 0.92)
refusal	True
pii	True
length	True (40–12000 chars)
language	True (en)

Rows removed by filters:

• none

Refusals held out of the dataset: 0.

Licence

apache-2.0

Generation

• Generated: 2026-09-08T17:10:18+00:00
• Dataset Genie version: 0.1.0
• Reproduce with genie run generation_config.yaml (the config sits next to this card).