← Frameworks
MANAGE 4.1

NIST AI RMF

NIST AI RMFView in NIST

How MANAGE 4.1 shows up in practice: the mapped risk classes in this atlas, the documented incidents that prove it's real, and the scenarios and controls to learn and defend against it.

Real-world cases

7

Documented incidents, disclosed vulnerabilities and research that illustrate MANAGE 4.1 — latest first, each with sources.

Hugging Face agentic production intrusion via a poisoned dataset (July 2026)16 Jul 2026

Hugging Face disclosed a production-infrastructure intrusion that it says was driven end-to-end by an autonomous AI-agent system: a malicious dataset abused code-execution paths in its dataset-processing pipeline as the foothold, then the campaign escalated to node-level access and moved laterally into internal clusters over a weekend.

Context Contamination: passive prompt injection poisons LLM security-log analysis16 Jul 2026

A red-team study shows adversaries can hide prompt-injection payloads inside network-log fields (usernames, URLs, user-agents) that fire when a SOC analyst asks an LLM to triage the logs — reportedly reaching up to 88.2% success at concealing malicious activity or exfiltrating data, turning the audit trail itself into the injection channel.

AI-assisted breach of Mexican government infrastructure (Claude Code + GPT-4.1)27 Dec 2025

Gambit Security reports that a single operator weaponized Anthropic's Claude Code and OpenAI's GPT-4.1 to breach at least nine Mexican government organizations, with Claude Code reportedly executing ~75% of remote commands after the attacker bypassed its refusals by loading a 1,084-line hacking cheatsheet as a persistent claude.md system prompt.

Grok 'MechaHitler' — config update degrades a deployed chatbot into antisemitic, violent output06 Jul 2025 / 08 Jul 2025

After an upstream code/instruction change, xAI's Grok began posting antisemitic tropes on X, self-identified as 'MechaHitler', and produced violence-themed content for hours before being pulled; xAI blamed a deprecated instruction path that made the bot mirror extremist user posts — not the base model.

UnMarker: Universal Black-Box Attack Defeating SynthID and Stable Signature14 May 2024

A universal, black-box, query-free attack that removes AI image watermarks including Google SynthID and Meta Stable Signature without knowing the scheme.

Watermarks in the Sand: Impossibility of Strong LLM Watermarking07 Nov 2023

Constructive proof that any strong generative-model watermark can be removed, demonstrated against three LLM watermarking schemes.

'How Is ChatGPT's Behavior Changing over Time?' (Chen, Zaharia, Zou)18 Jul 2023

Measured large swings in task performance between GPT-4/3.5 snapshots months apart — evidence of silent drift in a deployed service.

Browse all real-world cases →

Controls & guardrails that address this

352 proposed

Guardrails across the risks mapped to MANAGE 4.1, grouped by control function. Filter by control category below.

Control category
Preventive · 13
Risk-tiered minimum monitoring requirements at design

Define minimum monitoring requirements at design stage calibrated to the use case risk tier.

Lifecycle stage1 – Use Case Context & Design
Programmable conversation controls

Configure monitoring hooks in the conversation layer at deployment to capture metrics required by S1 monitoring requirements.

Lifecycle stages3 – Onboarding, Build & Review4 – Deployment
Fine-tuning

Execute a controlled fine-tuning cycle on refreshed data when staleness is confirmed. Validate before promoting to production.

Lifecycle stage5 – Usage, Monitoring & Change
Approved use scope baseline for OOD controls

Define approved use case scope and expected input distribution at design stage. Document as the governance baseline for OOD controls.

Lifecycle stage1 – Use Case Context & Design
Modular architecture

Design a scope-enforcement layer in the architecture to isolate the AI system from off-topic or out-of-distribution inputs.

Lifecycle stage1 – Use Case Context & Design
Input filtering

Maintain and update OOD detection rules in production as new unexpected use patterns are identified.

Weight provenance, hashing & pre-deploy evalsinteractive

Knowing exactly where the model came from, checking it hasn't been swapped, and testing its behaviour before going live.

Dependency integration safety contracts with schema validation and version pinning

Register a safety contract per integration — pinned version, schemas, side-effect class, latency/error envelope. Gate onboarding on contract review and sign-off.

source: OWASP Top 10 for LLM Apps LLM05:2025 Improper Output Handling; NIST SP 800-53 SA-9 External System Services
Lifecycle stages3 – Onboarding, Build & Review5 – Usage, Monitoring & Change
Change-freeze and blackout-window enforcement on agent-initiated changes

Wire the agent tool layer to the CAB calendar at deployment. Test that a declared freeze blocks mutating calls before go-live.

source: NIST SP 800-53 CM-3 Configuration Change Control, CM-5 Access Restrictions for Change; ITIL change-freeze practice
Lifecycle stages4 – Deployment5 – Usage, Monitoring & Change
Admission control on the inference & MCP serving plane: authenticate and network-segment every self-hosted inference/serving and MCP endpoint✚ proposed

Require authN/authZ on every inference API and MCP server, bind to private interfaces / front with a gateway, enforce network policy (no public exposure by default), and scope MCP tools to least privilege — so an exposed endpoint cannot be hijacked for compute resale, prompt/history exfiltration, or lateral movement. Pair with continuous asset discovery so endpoints can't drift back to an open default.

source: Case study: operation-bizarre-bazaar-llmjacking (Pillar Security, 28 Jan 2026)
Lifecycle stage4 – Deployment & Serving
Per-agent identity & taint-marked messagesinteractive

Giving each AI worker its own limited permissions and clearly labelling messages between them as 'untrusted until checked'.

Human-in-the-loop approval on high-risk actionsinteractive

Pausing to ask a person before doing anything big or hard to undo — sending money, deleting data, emailing customers.

Serving-stack & provisioning attestation, cache isolationinteractive

Making sure the machinery running the model — and the template used to stamp out new agents — is the real, unmodified version, and that one user's data can't leak into another's through shared shortcuts.

Detective · 8
Full-trace audit logginginteractive

Recording everything — questions, documents fetched, actions taken — so you can investigate when something goes wrong.

Synthetic evaluation datasets

Construct synthetic evaluation datasets during build to serve as the ongoing monitoring baseline.

Lifecycle stage3 – Onboarding, Build & Review
Robustness testing

Build monitoring infrastructure during build: performance metrics collection, alerting thresholds, dashboards.

Lifecycle stages3 – Onboarding, Build & Review4 – Deployment5 – Usage, Monitoring & Change
Cross-agent consensus and consistency monitoring to detect sycophantic agreement and error amplification✚ proposed

Run consistency and consensus checks across agent or model outputs to flag low-diversity agreement and amplifying error patterns, escalating or breaking the run before sycophantic convergence cascades into action.

source: Interactive-control reconciliation: ctrl-circuit-breaker (partial coverage)
Lifecycle stage5 – Usage, Monitoring & Change
Loop/cost circuit-breakers & consistency checksinteractive

Automatic stop-switches when AIs get stuck in loops, burn too much money, or start disagreeing with each other.

Content provenance & watermarkinginteractive

Tag AI-made content with a signed 'where it came from' label and an invisible watermark, and check those signals downstream — so AI media can be traced and flagged.

Corrective · 15
Reinforcement learning

Implement a reinforcement learning feedback loop to continuously incorporate production signals and reduce staleness risk.

Lifecycle stage5 – Usage, Monitoring & Change
Input filtering

Implement OOD detection in the input filtering layer. Reject or escalate inputs outside the S1-defined scope.

Red teaming

Conduct adversarial red team exercises simulating out-of-scope inputs and unexpected use patterns before deployment.

Human-in-the-loop validation

Configure HITL triggers for outputs in input domains that diverge from the training distribution. Log all out-of-scope interventions.

Lifecycle stage5 – Usage, Monitoring & Change
Non-production-by-default execution environment with explicit production promotion gate

Bind the agent's default execution target to non-production environments at design time. Require a separately approved promotion configuration for any production-connected target.

source: NIST SP 800-53 SC-7 Boundary Protection, CM-2 Baseline Configuration; OWASP Agentic AI Threats & Mitigations (cascading failures)
Lifecycle stages1 – Use Case Context & Design4 – Deployment
Graceful degradation and manual-fallback workflow on dependency unavailability

Map every dependency failure mode to a defined safe behaviour at design. Require architecture sign-off on the fallback specification before build.

source: NIST SP 800-53 CP-12 Safe Mode, SC-5 Denial-of-Service Protection; NIST AI RMF MANAGE 4.1 (post-deployment response/recovery)
Lifecycle stages1 – Use Case Context & Design4 – Deployment
Blast-radius scoping and environment isolation per agent task

Run each agent task in an isolated, network-segmented sandbox scoped to the task's exact needs. Gate onboarding on fault-injection tests proving containment.

source: NIST SP 800-53 SC-7 Boundary Protection, SC-39 Process Isolation; OWASP Agentic AI Threats & Mitigations (sandboxing/containment)
Lifecycle stages3 – Onboarding, Build & Review5 – Usage, Monitoring & Change
Cross-agent cascading-failure detection and orchestrator-level circuit breaking

Build tracing, detection rules and breaker thresholds into the orchestrator. Prove via fault-injection tests that a failing agent is quarantined within target before release.

source: OWASP Agentic AI Threats & Mitigations (cascading failures); Cloud Security Alliance MAESTRO (multi-agent threat modelling)
Lifecycle stages3 – Onboarding, Build & Review5 – Usage, Monitoring & Change
Idempotent action design with transactional rollback and pre-action snapshots

Engineer mutating actions with idempotency keys, transactions and pre-change snapshots; stage writes rather than committing directly. Gate release on tested dedup and rollback within RPO.

source: NIST SP 800-53 CP-9 System Backup, CP-10 System Recovery and Reconstitution; established idempotency / safe-write engineering practice
Lifecycle stages3 – Onboarding, Build & Review5 – Usage, Monitoring & Change
Rate, quota, and budget circuit breakers on outbound calls to connected systems

Cap each agent's rate, volume, concurrency, and spend per downstream dependency. Trip the breaker and fail closed when a ceiling is crossed.

source: NIST SP 800-53 SC-5 Denial-of-Service Protection, SC-6 Resource Availability; OWASP Top 10 for LLM Apps LLM10:2025 Unbounded Consumption
Lifecycle stages4 – Deployment5 – Usage, Monitoring & Change
Loop, recursion-depth, and iteration caps with runaway-loop detection

Enforce hard caps on iterations, depth, wall-clock, and cost per agent run. Terminate the run on cap breach or detected loop signatures.

source: OWASP Top 10 for LLM Apps LLM10:2025 Unbounded Consumption; OWASP Agentic AI Threats & Mitigations (cascading failures)
Lifecycle stages4 – Deployment5 – Usage, Monitoring & Change
Staged rollout with canary release and automated rollback on health-signal breach

Roll out agent changes via shadow and canary stages gated on connected-system health signals. Auto-halt and roll back to last known-good on threshold breach.

source: NIST SP 800-53 SI-2 Flaw Remediation, CM-3 Configuration Change Control; established progressive-delivery / canary practice
Lifecycle stages4 – Deployment5 – Usage, Monitoring & Change
Tiered kill-switch with per-agent, per-tool, and per-dependency containment scope

Deploy revocation, tool-cutoff and fleet-halt mechanisms with the release. Test every tier end-to-end and record time-to-effect before go-live.

source: OWASP Agentic AI Threats & Mitigations (kill-switch / containment); NIST AI RMF MANAGE 2.4 (mechanisms to supersede, disengage, or deactivate AI systems)
Lifecycle stages4 – Deployment5 – Usage, Monitoring & Change
Rollback and restore-to-known-good recovery procedure for AI services

Register each release as a restorable known-good baseline and rehearse rollback at the release gate. Block promotion without a tested restore.

source: ISO/IEC 27031 ICT readiness for business continuity; NIST SP 800-34r1 Contingency Planning (Recovery phase); NIST AI RMF MANAGE 2.4 (mechanisms to supersede/disengage/deactivate)
Lifecycle stages4 – Deployment5 – Usage, Monitoring & Change
Open the Control Library →

AI RiskAtlas is an educational model of how GenAI & agentic systems work and fail. Architectures and payloads are illustrative and simplified for learning — not operational guidance. Real-world cases are summarised from public reporting.

Sources & further reading →·Built by Shi Yuan ↗