← Frameworks
LLM08:2025

Vector and Embedding Weaknesses

How LLM08:2025 Vector and Embedding Weaknesses 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

38

Documented incidents, disclosed vulnerabilities and research that illustrate LLM08:2025 — latest first, each with sources.

CoSnitch: one-click exfiltration and persistent memory rules in Microsoft Copilot Personal (CVE-2026-24301)18 Aug 2026

Varonis Threat Labs reportedly chained flaws in consumer Microsoft Copilot Personal so a crafted link with a hidden autorun parameter fired prompt execution on a single click, exfiltrated data from connected OAuth apps, and planted persistent memory rules that survived password resets and device re-enrollment; reportedly patched Aug 2026 with no evidence of abuse.

Hanover Institute seeds question-shaped content to steer ChatGPT/Perplexity on Gaza16 Aug 2026

An investigation reportedly described a state-linked campaign publishing content phrased as chatbot questions (generative engine optimization) so that ChatGPT and Perplexity retrieve and cite it when answering neutral questions about Gaza and the IDF.

Encrypted chain-of-thought isn't private: stealing reasoning traces from frontier APIs10 Aug 2026

A paper reportedly showing that the encrypted reasoning envelopes returned by Anthropic/OpenAI/Google APIs are interchangeable across a provider's own models, so replaying one into a weaker sibling recovers the hidden chain-of-thought in plaintext, with PII and credentials extracted from scraped blocks.

LLM Heist: hijacking a LiteLLM gateway for traffic interception, key theft and forged tool-calls03 Aug 2026

Researcher Johann Rehberger reportedly showed an attacker with admin access to a LiteLLM AI gateway can weaponize the legitimate model-update endpoint and the callback system to reroute victim traffic, log backend provider API keys, and inject forged tool-calls into agent responses after inference, bypassing prompt-level defenses.

Azure DevOps MCP confused-deputy — hidden PR comments hijack AI review agents21 Jul 2026

Manifold Security reported that Microsoft's official Azure DevOps MCP server returns instructions planted in a hidden HTML comment inside a pull-request description — invisible in the web UI but returned verbatim by the API — so a victim's AI review agent, acting under the victim's credentials, follows the hidden orders and reaches data the attacker could not access directly.

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.

ServiceNow AI Platform pre-auth sandbox-escape RCE (CVE-2026-6875)13 Jul 2026

A reported CVSS 9.5 pre-authentication flaw in the ServiceNow AI Platform where unauthenticated endpoints feed attacker input into a query-filter operator that evaluates it as JavaScript, escalating via a script-loading gadget into full sandbox-escape code execution; reportedly patched 13 Jul 2026 with in-the-wild exploitation days later.

xAI Grok Build CLI — covert full-repo/secrets upload despite privacy opt-out12 Jul 2026 / 13 Jul 2026

A security researcher (Cereblab) captured xAI's Grok Build CLI silently uploading complete local Git repositories — untracked working files, full commit history, and unredacted secrets — to a Google Cloud Storage bucket, reportedly roughly 27,800x more data than the coding task needed, with the user-facing privacy toggle having no effect on the uploads.

mem0 agent-memory server: unauthenticated memory read/write + plaintext LLM-key disclosure (CVE-2026-59705 / CVE-2026-59706)07 Jul 2026

mem0's openmemory/api registered routers with no auth: an unauthenticated attacker could read/write/delete any user's stored memories (or globally pause memory for DoS), while a companion flaw exposed stored LLM API keys in plaintext and enabled SSRF to cloud metadata endpoints.

MemGhost: stealthy one-email memory injection in persistent personal agents06 Jul 2026

An academic paper reports MemGhost, a one-shot indirect-injection attack in which a single ordinary email causes a memory-enabled personal agent to silently write an attacker's false fact into long-term memory, conceal the write from its reply, and rely on the fiction in later sessions.

Amazon Q Developer auto-loads workspace MCP configs, enabling zero-click AWS credential theft (CVE-2026-12957)26 Jun 2026

Wiz Research found Amazon Q's VS Code extension auto-loaded MCP server definitions from a repo's .amazonq/mcp.json with no consent or workspace-trust prompt; opening a booby-trapped repository silently spawned attacker-controlled processes that inherited the developer's full environment and could stream live AWS session credentials out to an attacker.

+ 26 more via the mapped risk pages above.

Browse all real-world cases →

Controls & guardrails that address this

444 proposed

Guardrails across the risks mapped to LLM08:2025, grouped by control function. Filter by control category below.

Control category
Preventive · 24
Role-based access controls

Design strict RBAC on training data repositories at design stage. Define approved contributor list and approval workflow.

Lifecycle stages1 – Use Case Context & Design2 – Data Acquisition & Processing4 – Deployment
Input filtering

Apply anomaly detection on the training data ingestion pipeline to identify poisoned or tampered batches.

RAG / knowledge-base ingestion allow-listing with continuous index integrity re-validation

Define and approve the source allow-list and write-time scanning during build. Prove non-allow-listed and injection-bearing writes are rejected before go-live.

source: OWASP Top 10 for LLM Apps LLM04:2025 Data and Model Poisoning, LLM08:2025 Vector and Embedding Weaknesses; NIST SP 800-53 AC-3 / SI-7
Lifecycle stages3 – Onboarding, Build & Review5 – Usage, Monitoring & Change
Ingestion sanitisation & source allowlistinginteractive

Cleaning documents as they enter the library — stripping hidden text and active instructions — and only ingesting from trusted places.

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.

Approved storage location policy from collection

Establish data transfer and storage policy for AI training data. Enforce approved storage locations from point of collection.

Lifecycle stage2 – Data Acquisition & Processing
DLP controls in data acquisition environment

Implement DLP controls in the data acquisition environment to prevent unauthorised extraction or transfer of training data.

Lifecycle stage2 – Data Acquisition & Processing
Approval-gated data transfers from build environment

Enforce data handling policy in the build environment. Require explicit approval for any data transfers outside the environment.

Lifecycle stage3 – Onboarding, Build & Review
DLP controls confining build-environment training data

Configure DLP controls in the build environment to block training data from leaving approved boundaries.

Lifecycle stage3 – Onboarding, Build & Review
Privacy risk assessment and DPIA determination

Conduct a privacy risk assessment at use case design stage. Determine if a DPIA is required before data acquisition.

Lifecycle stage1 – Use Case Context & Design
Consent, minimisation, and anonymisation during acquisition

Apply S1-defined privacy controls during data acquisition: verify consent, minimise data, anonymise personal data.

Lifecycle stage2 – Data Acquisition & Processing
Validated anonymisation and masking before training

Apply anonymisation and masking controls to personal data before use in model training. Validate de-identification effectiveness.

Lifecycle stage2 – Data Acquisition & Processing
Privacy by Design via differential privacy

Apply Privacy by Design in model architecture using differential privacy or federated learning where technically feasible.

Lifecycle stage3 – Onboarding, Build & Review
Operational consent management and privacy notice

Publish the privacy notice and confirm consent management is operational before go-live.

Lifecycle stage4 – Deployment
Purpose-limitation enforcement on agent tool calls and cross-system data aggregation

Define and sign off a purpose-to-data-source matrix with lawful basis at intake. Make it the approved baseline for runtime enforcement.

source: NIST AI RMF MAP 1.1 / MANAGE 2.2 (context and intended purpose); NIST SP 800-53 AC-4 / AC-3 (purpose-based access enforcement)
Lifecycle stages1 – Use Case Context & Design5 – Usage, Monitoring & Change
Inference-time PII redaction and third-party LLM data-processing controls

Sign zero-retention/no-training terms with each model provider and obtain DPO sign-off on the data flow before enabling any endpoint.

source: OWASP Top 10 for LLM Apps LLM02:2025 Sensitive Information Disclosure; NIST SP 800-53 SC-8 / AC-4 (information flow enforcement)
Lifecycle stages3 – Onboarding, Build & Review4 – Deployment
Input/output filtering

Implement output filters to detect and suppress quasi-identifying attribute combinations in model responses.

Query-time access-control filtering of the retrieval/RAG corpus by caller entitlements (document-level ACL enforcement)

Propagate source ACLs and classification labels onto every chunk at ingestion. Reject documents whose entitlements cannot be resolved.

source: OWASP Top 10 for LLM Apps LLM02:2025 Sensitive Information Disclosure; NIST SP 800-53 AC-3 / AC-4 Information Flow Enforcement; OWASP Agentic AI Threats & Mitigations (privilege compromise)
Lifecycle stages2 – Data Acquisition & Processing3 – Onboarding, Build & Review5 – Usage, Monitoring & Change
Output-side DLP inspection with named-entity and PII redaction on the response path

Scan every model response inline with DLP before delivery; redact or block PII, PAN and MNPI matches. Keep the rule set version-controlled.

source: OWASP Top 10 for LLM Apps LLM02:2025 Sensitive Information Disclosure; NIST SP 800-53 SC-7(10) Prevent Exfiltration, SI-4
Lifecycle stages4 – Deployment5 – Usage, Monitoring & Change
Vet allowlisted egress destinations for server-side-fetch (SSRF) primitives; exclude or proxy-inspect any allowlisted service that can fetch arbitrary attacker-controlled URLs✚ proposed

An egress allowlist only contains exfiltration if no allowlisted destination can be coerced into fetching an attacker-controlled URL. Audit each allowlisted domain/endpoint for image-search / link-preview / URL-fetch features (SSRF proxies), and either remove them, pin them to fixed paths, or route them through an inspecting forward proxy. Pair with finishing output sanitization before render so no auto-fetch fires un-inspected.

source: Case study: searchleak-copilot (Varonis Threat Labs, CVE-2026-42824; reported by Microsoft as critical, mitigated server-side ~Jun 2026)
Lifecycle stage4 – Deployment & Serving
Egress allowlisting & DLP on tool argumentsinteractive

Controlling where the AI can send data, so secrets can't be quietly shipped to a stranger's address or website.

Per-user retrieval ACLsinteractive

Making sure the library only returns documents this particular user is allowed to see.

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 · 12
Vulnerability assessment

Conduct a data poisoning threat assessment at design stage. Identify likely attack vectors and assign risk ratings.

Red teaming

Simulate data poisoning attacks (backdoor, label flipping, gradient-based) to assess model resilience before deployment.

Cryptographic data provenance and signed dataset lineage (C2PA/in-toto attestations)

Verify a signed attestation and content hash on every dataset shard at ingestion. Reject unsigned or hash-mismatched data before it reaches the training pipeline.

source: MITRE ATLAS AML.M0007 (Sanitize Training Data), AML.M0014 (Verify ML Artifacts); NIST SP 800-53 SI-7 Software, Firmware, and Information Integrity, SR-4 Provenance
Lifecycle stages2 – Data Acquisition & Processing3 – Onboarding, Build & Review
Pre-deployment poisoning regression gate via canary backdoor probes and behavioral diff testing

Gate every model promotion on backdoor-trigger probes and a behavioral diff against the approved baseline. Block release on significant regressions or trigger-pattern anomalies.

source: MITRE ATLAS AML.M0014 (Verify ML Artifacts), AML.M0019 (Red Teaming); NIST AI RMF MANAGE 2.2 and MEASURE 2.7
Lifecycle stages3 – Onboarding, Build & Review5 – Usage, Monitoring & Change
Retrieval-time source-reputation weighting and coordinated-inauthentic-content detection for open-web / live-search assistants, with per-citation provenance shown to the user✚ proposed

For assistants that retrieve from the open web, rank and weight results by authenticated source reputation and independence — not just relevance / query-form match — so an anonymous, newly-registered, single-purpose site cannot become authoritative grounding. Run coordinated-inauthentic-content detection (look-alike site clusters, missing byline / legal entity, passages engineered for query-agnostic retrieval) and quarantine suspect sources. Surface per-citation provenance so users can see and discount low-trust sources. Does not defeat a well-resourced GEO campaign outright; it raises the cost and shrinks the yield.

source: Case study: hanover-institute-generative-engine-poisoning (per Politico investigation as reported by Arab News and Calcalist; FARA-disclosed funding/orchestration attributed to the reporting, not independently confirmed)
Lifecycle stage4 – Deployment & Serving
Provenance & content signinginteractive

Keeping a label on every document saying where it came from, so you can tell trusted company docs from random web text.

Real-time monitoring of anomalous data transfers

Monitor production for anomalous data transfers in real time. Alert on any transfer outside approved data flow boundaries.

Lifecycle stage5 – Usage, Monitoring & Change
Automated DSAR and right-to-erasure propagation across AI artefacts

Tag personal data with subject identifiers at ingestion and maintain an artefact inventory map of every store it reaches. Keep lineage current so erasure can propagate.

source: NIST AI RMF MANAGE 4.1 (post-deployment response); NIST SP 800-53 SI-12 Information Management and Retention, PT-2/PT-3 (personal data processing)
Lifecycle stages2 – Data Acquisition & Processing5 – Usage, Monitoring & Change
Canary-token and membership-inference red-team probes against training/fine-tuning data memorisation

Seed registered canary records into the fine-tuning corpus during data preparation. Control the seed manifest so canaries stay traceable and tamper-proof.

source: MITRE ATLAS AML.T0024 (Exfiltration via ML Inference API), AML.T0024.000 (Infer Training Data Membership); NIST AI RMF MEASURE 2.7
Lifecycle stages2 – Data Acquisition & Processing3 – Onboarding, Build & Review
Input guardrail / injection classifierinteractive

A screen that reads incoming messages and blocks obvious attacks or banned topics before the model sees them.

Full-trace audit logginginteractive

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

Corrective · 10
Penetration testing

Penetration test the training data pipeline to identify injection points and access control weaknesses.

Statistical anomaly and backdoor-trigger detection on ingested data (activation clustering / spectral signatures)

Scan every ingestion batch with spectral-signature and clustering detectors before training. Quarantine flagged clusters for human review against documented thresholds.

source: MITRE ATLAS AML.M0007 (Sanitize Training Data); OWASP Top 10 for LLM Apps LLM04:2025 Data and Model Poisoning; NIST AI RMF MEASURE 2.7
Lifecycle stages2 – Data Acquisition & Processing5 – Usage, Monitoring & Change
Bind long-term memory to the credential/session epoch: invalidate or force re-review of persisted memory on password reset, session revocation, or device re-enrollment✚ proposed

Tie the persistent-memory lifecycle to identity state so that standard remediation actually ends a compromise. On password reset, credential rotation, session revocation or device re-enrollment, invalidate (or quarantine for re-review) memory entries — especially entries whose provenance traces to summarised untrusted content — so a planted standing instruction cannot outlive the reset. Pair with write-path validation/provenance so instruction-shaped memory-writes from web content are caught on the way in.

source: Case study: cosnitch-copilot-personal-oneclick (Varonis Threat Labs, CVE-2026-24301; reportedly patched 18 Aug 2026, no evidence of abuse)
Lifecycle stage5 – Usage, Monitoring & Change
Runtime memory-poisoning drift detection and per-session memory quarantine/rollback✚ proposed

Continuously correlate live agent-memory writes against output behaviour to flag drift, then quarantine and roll back the suspected-poisoned memory record across all affected sessions.

source: Interactive-control reconciliation: ctrl-memory-quarantine (partial coverage)
Lifecycle stage5 – Usage, Monitoring & Change
Production privacy incident monitoring and regulator notification

Monitor for privacy incidents in production including personal data appearing in outputs. Notify regulators within required timeframes.

Lifecycle stage5 – Usage, Monitoring & Change
Privacy hygiene for agent memory and RAG/vector stores (retention, scoping, erasure of embeddings)

Tag every memory and vector record with subject-id and retention class; partition stores per tenant/user. Prove the erasure and isolation paths in testing before release.

source: OWASP Agentic AI Threats & Mitigations (memory/knowledge-base privacy); NIST SP 800-53 SI-12 Information Management and Retention
Lifecycle stages3 – Onboarding, Build & Review5 – Usage, Monitoring & Change
Red teaming

Test de-identification approach against known re-identification attacks (quasi-identifier linkage, singling-out). Remediate if risk is high.

Vulnerability assessment

Conduct periodic data leakage audits including training data memorisation testing. Escalate confirmed leakage incidents to PDPA notification process.

Forensic evidence preservation and incident logging

Implement tamper-evident capture of prompts, outputs, and version state during build. Verify a full incident timeline can be reconstructed before go-live.

source: NIST SP 800-86 Guide to Integrating Forensic Techniques into Incident Response; ISO/IEC 27037 evidence handling; NIST SP 800-61r2 (Detection & Analysis – evidence handling)
Lifecycle stages3 – Onboarding, Build & Review5 – Usage, Monitoring & Change
Egress allow-listing and tool-call sandboxing to block exfiltration of injected/sensitive data by agents

Run agent tool calls in a network-restricted sandbox behind a deny-by-default egress allow-list. Require security approval for any destination added.

source: OWASP Top 10 for LLM Apps LLM02:2025 Sensitive Information Disclosure; OWASP Agentic AI Threats & Mitigations (tool-misuse / exfiltration); NIST SP 800-53 SC-7 Boundary Protection / AC-4
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 ↗