BD ciso-agent-security
AI agent cybersecurity skill implementing MITRE ATLAS, OWASP Top 10 for LLM and Agentic Applications, CSA MAESTRO, NIST AI RMF, and Gray Swan frameworks. Red team patrol procedures, posture scoring, quarantine enforcement, and patch standards for autonomous AI agent systems.
As a process D 49/100 · Unfinished process — weak spots: result and completion, inputs and preconditions, failures and branches
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 · 9
✓ No critical or high findings
Medium and low: 9
-
low Risky intent
intent-offensive-securityREADME.md:3Offensive-security / dual-use content (legitimate for authorised testing; review intended use)AI agent cybersecurity skill for autonomous agent systems. Implements six industry-standard security frameworks for red team patrols, vulnerability assessments, posture scoring, quarantine enforcement
-
low Risky intent
intent-offensive-securityREADME.md:9Offensive-security / dual-use content (legitimate for authorised testing; review intended use)| **MITRE ATLAS** | Adversary tactics, techniques, and procedures (TTPs) for AI systems. Primary red team attack pattern reference. |
-
low Risky intent
intent-offensive-securityskill-card.md:2Offensive-security / dual-use content (legitimate for authorised testing; review intended use)AI agent cybersecurity skill for autonomous agent systems that implements MITRE ATLAS, OWASP Top 10 for LLM and Agentic Applications, CSA MAESTRO, NIST AI RMF, and Gray Swan frameworks for red team pa
-
low Risky intent
intent-offensive-securityskill-card.md:14Offensive-security / dual-use content (legitimate for authorised testing; review intended use)Security engineers, CISO teams, and agent developers use this skill to guide AI agent red team patrols, vulnerability assessments, posture scoring, quarantine decisions, and patch recommendations agai
-
low Risky intent
intent-offensive-securitySKILL.md:3Offensive-security / dual-use content (legitimate for authorised testing; review intended use)description: AI agent cybersecurity skill implementing MITRE ATLAS, OWASP Top 10 for LLM and Agentic Applications, CSA MAESTRO, NIST AI RMF, and Gray Swan frameworks. Red team patrol procedures, postu
-
low Risky intent
intent-offensive-securitySKILL.md:20Offensive-security / dual-use content (legitimate for authorised testing; review intended use)# CISO Security Skill -- AI Agent Red Teaming and Defense
-
low Risky intent
intent-offensive-securitySKILL.md:24Offensive-security / dual-use content (legitimate for authorised testing; review intended use) (detector / deny-list definition)This skill defines the frameworks, methods, and official sources the CISO agent uses when conducting security patrols, red team testing, vulnerability assessments, and posture scoring across the agent
detector -
low Risky intent
intent-offensive-securitySKILL.md:36Offensive-security / dual-use content (legitimate for authorised testing; review intended use)**Role:** Primary red team attack pattern reference. Use ATLAS to identify adversary tactics, techniques, and procedures (TTPs) specific to AI systems. All patrol test cases should map to ATLAS techni
-
low Risky intent
intent-offensive-securitySKILL.md:183Offensive-security / dual-use content (legitimate for authorised testing; review intended use) (detector / deny-list definition)Run targeted red team tests using ATLAS technique patterns relevant to the agent's role:
detector
Files scanned: 3. Evidence is masked. Grey chips explain why severity was lowered.
Against the Agent Skills spec
- warning
description-no-whendescription does not say WHEN to use the skill (no "use when")
Process rating: all ten parameters 49/100
- 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
- 30Running it twice. 1 mutating operations with no state check
- 50When it triggers. No condition that starts the skill
- 60Tools and files. Uses tools (bash, web) that frontmatter does not declare
- 100Steps. 85 steps
- 100Consistency. Name and required fields are in place
- 100Execution cost. Instruction body is 2343 tokens
- 100Progress reporting. Reports progress
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
- +4No input/output examples
- +2Single-language instructions
- +3Description length 275: enough signal without eating the budget
- +4Structure: 20 headings
- +3Step-by-step instructions: 85 items
- +1License stated
Quality base 70; lint remarks subtract, signals add up to 100. Result: 69.
External checks
ClawHub: clean
This is a disclosed, instruction-only security assessment skill, with operational cautions around any real quarantine or patch authority granted to the receiving agent.
LLM: benign (high) · VirusTotal: · 29 May 2026