OWASP AIVSS (AI Vulnerability Scoring System) scores security vulnerabilities in agentic AI systems. It extends CVSS v4.0 with 10 factors that measure how agent autonomy, tool use, and persistence amplify the blast radius of technical vulnerabilities.

Version 0.8 was released March 19, 2026. The review board includes Jason Clinton (Deputy CISO, Anthropic), Rob Joyce (ex-NSA, advisor to OpenAI), and Martin Stanley (NIST AI RMF lead). RSAC 2026 presentation: March 23.

The Mitigation Factor

AIVSS includes a Mitigation Factor in its scoring formula. This is where pipelock fits.

Mitigation StrengthFactorWhat It Means
No/Weak Mitigation1.00Mitigations absent or ineffective
Partial Mitigation0.83Some mitigations exist but incomplete or unvalidated
Strong Mitigation0.67Effective mitigations in place, validated, consistently enforceable (fail-closed where applicable)

Pipelock can support a strong-mitigation assessment when deployed and validated in your environment. The 0.67 value is provisional in v0.8 and subject to calibration. Here’s how pipelock aligns with the criteria:

Deploying pipelock can reduce an agent’s AIVSS vulnerability score by up to 33%.

Coverage Mapping: 10 OWASP Agentic AI Core Security Risks

AIVSS defines 10 broad risk categories. These are system-level vulnerability classes, not individual controls. Pipelock mitigates specific scenarios within each category at the network layer. Some sub-risks (like deepfake identity abuse or compromised open-source libraries) fall outside pipelock’s scope and require other defenses.

Note: AIVSS distinguishes between “10 Core Security Risks” (the vulnerability taxonomy) and “10 Risk Amplification Factors” (the scoring dimensions). This page maps to the risks. The amplification factors are used in the scoring formula below.

Risk 1: Agentic AI Tool Misuse

The AIVSS spec explicitly calls out two controls as missing risks under Tool Misuse:

“Lack of Data Inspection: Absence of robust Data Loss Prevention (DLP) and inspection mechanisms creates significant risk of sensitive data exfiltration.”

“Absence or Inadequacy of Emergency Termination Controls (Kill Switch Mechanism): Without network-isolated or infrastructure-level kill switches, compromised agents cannot be reliably stopped.”

Pipelock’s response:

Sub-riskPipelock Feature
Tool squatting / poisoned descriptionsMCP tool scanning detects hidden instructions in tool metadata
Insecure tool invocationMCP tool policy with shell normalization (blocks command injection)
Compromised tool usageMCP input scanning catches DLP violations in tool arguments
Emergent capability triggeringTool drift detection flags schema changes (rug-pull detection)
Lack of authoritative inventorySession binding pins tool inventory per session (max 10,000 tools)
Lack of DLP inspection46 DLP patterns + entropy analysis + env variable leak detection
Absent kill switchOR-composed kill switch: config, API, SIGUSR1, sentinel file
Malicious MCP serverTool scanning on first connect + drift detection on subsequent calls

Risk 2: Agent Access Control Violation

The spec describes a “Rogue Proxy Hijacking (AgentSmith)” attack where a misconfigured proxy intercepts agent traffic.

Pipelock’s response: Pipelock IS the proxy, by design. Capability separation architecture ensures the agent (which has secrets) cannot access the network directly. All traffic routes through pipelock (which has no agent secrets but has network access). This is the opposite of a rogue proxy: it’s an intentional, configured security boundary.

Sub-riskPipelock Feature
Credential/token mismanagementDLP scans all outbound traffic for leaked credentials
Memory-based data leakageResponse scanning detects injection payloads in tool results
Cross-context privilege bleedPer-agent identity and isolated configs (Pro tier)

Risk 3: Agent Cascading Failures

Pipelock’s response:

Sub-riskPipelock Feature
Cross-system exploitationDomain blocklist + SSRF protection limit lateral movement
Impact amplificationRate limiting per destination + data budget per domain
SaaS-to-SaaS pivotingAll outbound requests scanned regardless of destination
Hallucination propagationResponse scanning catches injected content in tool results

Kill switch provides immediate shutdown across all traffic. Fail-closed defaults prevent partial failure states.

Risk 4: Agent Orchestration and Multi-Agent Exploitation

Pipelock’s response:

Sub-riskPipelock Feature
Inter-agent communication exploitationMCP input scanning on all tool calls
Shared knowledge poisoningResponse scanning detects injected instructions
Session fixation / replayMCP session binding pins tool inventory per session
Capability drift / schema mismatchTool drift detection flags changed tool definitions
Coordination protocol manipulationTool chain detection catches suspicious call sequences

Risk 5: Agent Identity Impersonation

Pipelock’s response:

Sub-riskPipelock Feature
Agent impersonationSession binding creates per-session tool fingerprints
Shared identity poolsPer-agent identity with isolated configs, budgets, CIDR matching
Agent-in-the-middleProxy architecture places pipelock at the network boundary

Risk 6: Agent Memory and Context Manipulation

The spec describes “Invisible Context Injection” using Unicode characters and “RAG Data Poisoning” via prompt injections in retrieved content.

Pipelock’s response:

Sub-riskPipelock Feature
Context manipulationResponse scanning with Unicode normalization (3 passes)
Invisible context injectionText normalizer strips zero-width characters, homoglyphs
RAG data poisoningResponse scanning detects injection in fetched/retrieved content
Cross-session data leakageDLP scanning runs on every request regardless of session state

Risk 7: Insecure Agent Critical Systems Interaction

The spec includes a detailed SSRF attack scenario where an agent’s fetch_internal_doc(url) tool is tricked into accessing http://10.0.1.20/status (internal admin dashboard), then used as a proxy to scan the internal network.

Pipelock’s response: This is exactly what pipelock’s SSRF layer prevents.

Sub-riskPipelock Feature
SSRF into internal systemsLayer 6: private IP blocking, metadata endpoint blocking, DNS rebinding protection
CI/CD pipeline tamperingDomain blocklist + tool policy rules for destructive operations
Multi-system simultaneous manipulationRate limiting + data budget constrain blast radius
Feedback loop exploitationRate limiting prevents resource exhaustion loops

DLP runs before DNS resolution (layers 2-3 before layer 6), preventing secret exfiltration via DNS queries.

Risk 8: Agent Supply Chain and Dependency Risk

The spec describes “Malicious MCP Honeypots” that offer free tools while secretly scraping data and injecting instructions.

Pipelock’s response:

Sub-riskPipelock Feature
Malicious MCP server dependencyTool scanning detects hidden instructions in tool metadata
Tool metadata manipulationRug-pull drift detection flags schema changes after initial connect
Package hallucinationsTool policy blocks unauthorized package installation commands
Capability creepSession binding detects new tools appearing mid-session
Signed rule bundlesCommunity detection patterns with Ed25519 signature verification

Risk 9: Agent Untraceability

Pipelock’s response:

Sub-riskPipelock Feature
Log tampering or absencePrometheus metrics + webhook/syslog event emission
Loss of chain-of-actionPer-request logging with agent identity, destination, action, bytes
Deceptive MCP tool invocationTool scanning flags tools with suspicious descriptions
Cross-domain attribution failurePer-agent identity tags every request with the originating agent

Event emission is fire-and-forget (async buffered channel for webhook, synchronous UDP for syslog). Neither blocks the proxy.

Risk 10: Agent Goal and Instruction Manipulation

This is the prompt injection risk. The spec covers direct injection, indirect injection via RAG, and gradual goal steering.

Pipelock’s response:

Sub-riskPipelock Feature
Direct instruction injectionResponse scanning with keyword-gated regex on all 3 normalization passes
Indirect instruction injection (RAG)Fetched content scanned before delivery to agent
Resource exhaustion via goal loopingRate limiting + data budget prevent DoS
Dynamic goal steeringAdaptive enforcement escalates restrictions based on behavior score

What AIVSS Measures vs. What Pipelock Does

AIVSS scores vulnerabilities. Pipelock prevents exploitation. The two are complementary:

AIVSS Scoring Formula

The full AIVSS v0.8 formula:

Risk_Gap    = 10 - CVSS_Base
AARS        = Risk_Gap * (Factor_Sum / 10) * Threat_Multiplier
AIVSS       = (CVSS_Base + AARS) * Mitigation_Factor

Where AARS (Agentic AI Risk Score) is the uplift from agentic capabilities, Factor_Sum is the sum of all 10 amplification factors (0.0-10.0), and Threat_Multiplier defaults to 0.97.

Hypothetical example (using the spec’s Tool Misuse scenario):

Without pipelock (Mitigation_Factor = 1.0): CVSS 9.4 + AARS 0.5 = 9.9 (Critical).

With pipelock (Mitigation_Factor = 0.67, if validated as strong mitigation): (9.4 + 0.5) * 0.67 = 6.6 (Medium).

Same vulnerability, same agentic context. The difference is whether validated mitigations are in place.

References