BD BloodHound Narrator
Turn BloodHound attack path exports into dual-layer security reports — CISO executive prose on top, technical remediation playbook below. Automates Active Directory audit reporting for pentesters, blue teams, and security consultants. Supports DCSync, Kerberoasting, AD CS abuse (ESC1/3/4/6/7/9/10/13, Golden Certificate), resource-based constrained delegation, shadow credentials, LAPS and gMSA password reads, ACL abuse, GPO takeover, and lateral movement paths. Pure local PowerShell — no API calls, no data leaves your machine. Air-gap compatible.
As a process D 41/100 · Unfinished process — weak spots: result and completion, when it triggers, inputs and preconditions
IntegrationData and analyticsInfrastructureSecuritytype and topics are labelled automatically from the skill text
How to improve
- Say in the description WHEN to use the skill ("use when…", example requests): that is the agent's main cue.
For the model run — optional
- Your own cases (evals/evals.json, 4–6 real requests with expected answers): the full check would then run those instead of a model-drafted suite.
- A spec.yaml with trigger phrases and assertions — a behaviour contract for CI; `skilltest init` writes a template.
Guard findings · 14
✓ No critical or high findings
Medium and low: 14
-
low Risky intent
intent-offensive-securityCHANGELOG.md:20Offensive-security / dual-use content (legitimate for authorised testing; review intended use)- **Additional lateral movement edges** — `CanRDP`, `CanPSRemote`, `ExecuteDCOM`, `SQLAdmin`, `WriteSPN`, `WriteAccountRestrictions`.
-
low Risky intent
intent-offensive-securityCHANGELOG.md:30Offensive-security / dual-use content (legitimate for authorised testing; review intended use)- Lateral movement rebalanced: `AdminTo` and `HasSession` now score 3 each individually, and the combination bonus dropped from 10 to 5 to avoid double counting.
-
low Risky intent
intent-offensive-securityscripts/lib/NarrativeTemplates.txt:87Offensive-security / dual-use content (legitimate for authorised testing; review intended use) (quoted — discussed, not commanded)$lines.Add("*This path enables significant lateral movement or privilege escalation that could compromise sensitive business systems and data.*`n")quoted -
low Risky intent
intent-offensive-securityscripts/lib/NarrativeTemplates.txt:152Offensive-security / dual-use content (legitimate for authorised testing; review intended use) (detector / deny-list definition)'Deploy Credential Guard on endpoints to prevent credential harvesting.'
detector -
low Risky intent
intent-offensive-securityscripts/lib/NarrativeTemplates.txt:211Offensive-security / dual-use content (legitimate for authorised testing; review intended use) (detector / deny-list definition)'GenericWrite on a user permits SPN injection (targeted Kerberoasting) and shadow-credential attacks — treat as privilege escalation, not a write permission.'
detector -
low Risky intent
intent-offensive-securityscripts/lib/NarrativeTemplates.txt:421Offensive-security / dual-use content (legitimate for authorised testing; review intended use)# ── v1.0.2: lateral movement ──
-
low Risky intent
intent-offensive-securityscripts/lib/NarrativeTemplates.txt:431Offensive-security / dual-use content (legitimate for authorised testing; review intended use) (quoted — discussed, not commanded)title = 'Break the Lateral Movement Chain'
quoted -
low Risky intent
intent-offensive-securityscripts/lib/SeverityClassifier.txt:103Offensive-security / dual-use content (legitimate for authorised testing; review intended use)# ── Lateral movement ──
-
low Risky intent
intent-offensive-securityscripts/lib/SeverityClassifier.txt:386Offensive-security / dual-use content (legitimate for authorised testing; review intended use)# ── Factor: lateral movement chain ──
-
low Risky intent
intent-offensive-securityscripts/lib/SeverityClassifier.txt:389Offensive-security / dual-use content (legitimate for authorised testing; review intended use) (quoted — discussed, not commanded)$reasons.Add('Lateral movement chain: local admin + session hijack enables credential theft')quoted -
low Risky intent
intent-offensive-securitySKILL.md:3Offensive-security / dual-use content (legitimate for authorised testing; review intended use)description: Turn BloodHound attack path exports into dual-layer security reports — CISO executive prose on top, technical remediation playbook below. Automates Active Directory audit reporting for pe
-
low Risky intent
intent-offensive-securitySKILL.md:21Offensive-security / dual-use content (legitimate for authorised testing; review intended use)**Detected attack patterns:** DCSync and directory replication rights, Kerberoasting, unconstrained delegation, resource-based constrained delegation (RBCD), shadow credentials (AddKeyCredentialLink),
-
low Risky intent
intent-offensive-securitySKILL.md:88Offensive-security / dual-use content (legitimate for authorised testing; review intended use)| Lateral movement chain (AdminTo + HasSession) | +5 |
-
low Risky intent
intent-offensive-securitySKILL.md:154Offensive-security / dual-use content (legitimate for authorised testing; review intended use)- Post-pentest AD audit reporting
Files scanned: 8. Evidence is masked. Grey chips explain why severity was lowered.
Against the Agent Skills spec
- warning
name-formatname should be kebab-case (lowercase letters, digits, hyphens) - warning
description-no-whendescription does not say WHEN to use the skill (no "use when") - note
frontmatter-keyunknown frontmatter key "bins"
Process rating: all ten parameters 41/100
- 0Result and completion. Does not say what the result is
- 0Inputs and preconditions. Does not say what the process needs to start
- 0Failures and branches. Linear process with no failure handling
- 0Progress reporting. Says nothing while it works
- 20When it triggers. No condition that starts the skill
- 40Consistency. Frontmatter name (BloodHound Narrator) differs from the folder (bloodhound-narrator)
- 60Tools and files. Uses tools (node) that frontmatter does not declare
- 100Steps. 18 steps
- 100Execution cost. Instruction body is 1739 tokens
- 100Running it twice. Mutating operations check current state
Everything here is measured from the skill text rather than judged by a model, so the numbers are checkable. A parameter weighs more when it is a more common reason for the process to stall.
Quality signals
- +5Description has no quoted example phrases that should trigger the skill
- +4Description does not say when NOT to use the skill (false activations)
- +3Output format is not stated: the model decides each time
- +2Single-language instructions
- +3Description length 551: enough signal without eating the budget
- +4Structure: 12 headings
- +3Step-by-step instructions: 18 items
- +4Has examples (3 code blocks)
- +3All 1 scripts are documented
- +1License stated
Quality base 70; lint remarks subtract, signals add up to 100. Result: 70.
External checks
ClawHub: clean
BloodHound Narrator is a coherent local reporting tool that reads user-provided BloodHound JSON and writes a Markdown report without network use or persistence.
LLM: benign (high) · VirusTotal: · 26 Jul 2026