KC1 · Language Models (LLMs)
Small-Language Models (SLMs) (KC1.3)
Language models with fewer parameters, trained on smaller, focused datasets, designed for specific tasks or use cases. Characterized by smaller weight space, parameter size, and context window compared to LLMs.
- 5 threats
- 12 controls
- 0 architectures
Threats
T5 Cascading Hallucination Attacks
Foundation models generate incorrect information that propagates through the system, affecting reasoning quality and being stored in memory across sessions or agents.
T6 Intent Breaking & Goal Manipulation
Attacks that manipulate the agent's core decision-making to achieve unauthorized goals, breaking control flow, abusing shared context, or interfering with isolated data.
T7 Misaligned & Deceptive Behaviors
Model alignment issues leading to unintended behaviors that impact users, organizations, or broader populations through subtle reasoning or tool usage misalignments.
T15 Human Manipulation
Models exploit human trust to manipulate users, leveraging reasoning capabilities or operational access to influence human decisions.
T17 Supply Chain Compromise
Vulnerable, malicious or outdated components enter the agent through models, libraries, tools, plugins, prompt templates, build environments or framework updates.
Controls
- M4 Prompt Hardening & Jailbreak Prevention: Implement multiple layers of prompt security including structural defenses, behavioral constraints, runtime detection, and advanced techniques to prevent prompt injection, model manipulation, and jailbreaking attempts.
- M5 Multi-Stage Reasoning Validation: Implement comprehensive validation of agent reasoning processes using multiple checkpoints, external validators, consistency checks, and fact verification to ensure goal alignment, prevent manipulation, and detect hallucinations.
- M6 Comprehensive Security Monitoring & Auditing: Implement end-to-end observability with security-focused monitoring, anomaly detection, tamper-evident logging, and advanced threat hunting to ensure accountability, enable incident response, and detect rogue agents.
- M8 Adaptive Human-in-the-Loop Controls: Implement intelligent human oversight with risk-based approval workflows, clear decision interfaces, anti-fatigue mechanisms, and adaptive trust systems to prevent approval overwhelm, human attacks, and manipulation while maintaining effective oversight.
- M11 Content Security & Output Filtering: Implement comprehensive filtering and validation of agent outputs to prevent harmful content generation, data leakage, manipulation attempts, and ensure compliance with security policies and regulatory requirements.
- M12 Secure Model Training & Validation: Comprehensive security practices for model training pipelines including data integrity verification, training environment hardening, model validation against adversarial inputs, and ensuring reproducibility of training processes.
- M13 Secure Development Lifecycle Integration: Embed security practices throughout the AI and agent lifecycle with security design reviews, automated testing gates, dependency scanning, and continuous compliance checks.
- M16 Behavioral Monitoring & Ethics Framework: Implement comprehensive behavioral analysis, ethics enforcement, and bias detection to identify misaligned behaviors, prevent human manipulation, and ensure ethical agent operations while maintaining transparency and accountability.
- M17 Supply Chain Security & Provenance: Implement comprehensive supply chain security including model provenance verification, dependency management, secure development environments, and third-party component validation to prevent supply chain attacks and ensure component integrity.
- M18 Incident Response & Recovery Framework: Implement comprehensive incident response capabilities specifically designed for agentic AI systems, including automated containment, forensic analysis, recovery procedures, and lessons learned integration to rapidly respond to security incidents and minimize impact.
- M19 MCP Server Hardening & Tool Provenance: Treat every MCP server, A2A peer and tool definition as an untrusted, versioned dependency: admit only allow-listed and verified servers, authenticate and authorize every protocol request, validate and screen tool metadata and results before they reach the model, and require re-approval when a tool definition changes.
- M20 Agent Supply-Chain Integrity: Extend software supply-chain controls to everything an agent loads at build or run time: models and adapters, prompt templates, agent cards, tool and MCP server definitions, plugins and framework updates.
Source: OWASP Securing Agentic Applications Guide (KC1-KC6 sub-components) 1.0 (Jul 2025) · retrieved 2026-11