{"items":[{"id":"NVI-00012185","kind":"vulnerability","title":"vLLM: Mirrored multimodal IPC caches desync after a rejected request — a later request reusing the same media hash trips a receiver assertion in the engine core","revision":2,"unknowns":["Customer inventory, deployment, reachability and exploitation have not been inspected."],"lifecycle":"active","identifiers":["GHSA-ph3r-5jfg-f84f","CVE-2026-105753"],"plain_summary":"GitHub lists this reviewed advisory for vllm. Published version ranges and source prerequisites are shown below; this is not an inspection of a visitor’s application.","review_status":"source_listed","control_topics":[],"public_actions":[{"summary":"Identify the exact component and deployed version, examine the linked publisher’s conditions, and retain evidence of any mitigation.","source_ids":["sr-8b42e6579c1f15adc8652f1b0b00b272"],"action_type":"investigate","limitations":"Confirm current vendor support, compatibility and deployment evidence. This report has not tested a fix in your system.","current_supported_target":null}],"schema_version":"1.1.0","source_metrics":[{"kind":"CVSS","as_of":"2026-10-06T00:02:07.000Z","value":"6.5 medium","meaning":"GitHub source severity; exploitability and local consequence require separate evidence.","source_id":"sr-c0e4bae2f1c07777bce5b004f4a9661f"},{"kind":"EPSS","as_of":"2026-10-31T12:00:27.000Z","value":"0.427%","meaning":"FIRST EPSS model v2026.06.15; exploitation probability over 30 days, percentile 0.37025. Missing scores are unknown.","source_id":"sr-8b42e6579c1f15adc8652f1b0b00b272"}],"why_it_matters":"Published affected context helps an owner decide what evidence to gather. This report has not examined your application.","related_sources":[{"id":"sr-c0e4bae2f1c07777bce5b004f4a9661f","url":"https://github.com/advisories/GHSA-ph3r-5jfg-f84f","basis":"vendor_advisory","notes":"GitHub Advisory Database; CC-BY-4.0. Each advisory links its original attribution. Null normalizes fields.","title":"GHSA-ph3r-5jfg-f84f","publisher":"GitHub Advisory Database","observed_at":"2026-11-01T00:59:16.959Z"},{"id":"sr-8b42e6579c1f15adc8652f1b0b00b272","url":"https://www.first.org/epss/data.html","basis":"official_catalog","notes":"FIRST EPSS; probability of exploitation in the next 30 days, not local compromise. Dataset date/model retained; FIRST services terms apply.","title":"CVE-2026-105753","publisher":"FIRST EPSS","observed_at":"2026-11-01T01:39:43.064Z"}],"affected_context":[{"certainty":"source_listed","source_ids":["sr-c0e4bae2f1c07777bce5b004f4a9661f"],"technology":"vllm","known_exclusions":[],"conditions_summary":"## Affected\n\n- **Ecosystem / package:** pip / `vllm`\n- **Affected versions:** vLLM ≤ 0.25.1 (confirmed on 0.25.1, commit `752a3a504485`). The lower bound predates 0.25.1; maintainers can confirm how far back the mirrored sender/receiver cache protocol reaches.\n\n## Summary\n\nvLLM's default multimodal cache (`mm_processor_cache_type=\"lru\"`) mirrors state across two processes: the frontend (P0) holds only metadata (`MultiModalProcessorSenderCache`) while the engine core (P1) holds the real payload (`MultiModalReceiverCache`). The design invariant is that `get_and_update()` runs on P0 and P1 in lockstep for every request, so eviction order stays mirrored and P0 can answer \"is this cached in P1?\" without talking to P1.\n\nThat invariant breaks when a request is **rejected after P0 has rendered and hashed the multimodal input** (populating the P0 cache) **but before P1 receives the item** — for example, an oversized chat prompt rejected on `max_model_len` *after* rendering. P0 now believes the media is cached while P1 never got it. A later request reusing the same media hash gets a P0 hit, so P0 sends `Non","public_affected_ranges":["< 0.28.0"]}],"null_published_at":"2026-11-04T02:53:12.380Z","source_updated_at":"2026-10-06T00:02:07.000Z","source_disclosed_at":"2026-10-06T00:02:05.000Z","is_synthetic_fixture":false,"null_last_reviewed_at":null,"public_change_summary":"Dated EPSS estimate updated; this is not a new disclosure or proof of exploitation.","null_first_observed_at":"2026-11-01T00:59:16.959Z"},{"id":"NVI-00012185","kind":"vulnerability","title":"vLLM: Mirrored multimodal IPC caches desync after a rejected request — a later request reusing the same media hash trips a receiver assertion in the engine core","revision":1,"unknowns":["Customer inventory, deployment, reachability and exploitation have not been inspected."],"lifecycle":"active","identifiers":["GHSA-ph3r-5jfg-f84f","CVE-2026-105753"],"plain_summary":"GitHub lists this reviewed advisory for vllm. Published version ranges and source prerequisites are shown below; this is not an inspection of a visitor’s application.","review_status":"source_listed","control_topics":[],"public_actions":[{"summary":"Identify the exact component and deployed version, examine the linked publisher’s conditions, and retain evidence of any mitigation.","source_ids":["sr-c0e4bae2f1c07777bce5b004f4a9661f"],"action_type":"investigate","limitations":"Confirm current vendor support, compatibility and deployment evidence. This report has not tested a fix in your system.","current_supported_target":null}],"schema_version":"1.1.0","source_metrics":[{"kind":"CVSS","as_of":"2026-10-06T00:02:07.000Z","value":"6.5 medium","meaning":"GitHub source severity; exploitability and local consequence require separate evidence.","source_id":"sr-c0e4bae2f1c07777bce5b004f4a9661f"}],"why_it_matters":"Published affected context helps an owner decide what evidence to gather. This report has not examined your application.","related_sources":[{"id":"sr-c0e4bae2f1c07777bce5b004f4a9661f","url":"https://github.com/advisories/GHSA-ph3r-5jfg-f84f","basis":"vendor_advisory","notes":"GitHub Advisory Database; CC-BY-4.0. Each advisory links its original attribution. Null normalizes fields.","title":"GHSA-ph3r-5jfg-f84f","publisher":"GitHub Advisory Database","observed_at":"2026-11-01T00:59:16.959Z"}],"affected_context":[{"certainty":"source_listed","source_ids":["sr-c0e4bae2f1c07777bce5b004f4a9661f"],"technology":"vllm","known_exclusions":[],"conditions_summary":"## Affected\n\n- **Ecosystem / package:** pip / `vllm`\n- **Affected versions:** vLLM ≤ 0.25.1 (confirmed on 0.25.1, commit `752a3a504485`). The lower bound predates 0.25.1; maintainers can confirm how far back the mirrored sender/receiver cache protocol reaches.\n\n## Summary\n\nvLLM's default multimodal cache (`mm_processor_cache_type=\"lru\"`) mirrors state across two processes: the frontend (P0) holds only metadata (`MultiModalProcessorSenderCache`) while the engine core (P1) holds the real payload (`MultiModalReceiverCache`). The design invariant is that `get_and_update()` runs on P0 and P1 in lockstep for every request, so eviction order stays mirrored and P0 can answer \"is this cached in P1?\" without talking to P1.\n\nThat invariant breaks when a request is **rejected after P0 has rendered and hashed the multimodal input** (populating the P0 cache) **but before P1 receives the item** — for example, an oversized chat prompt rejected on `max_model_len` *after* rendering. P0 now believes the media is cached while P1 never got it. A later request reusing the same media hash gets a P0 hit, so P0 sends `Non","public_affected_ranges":["< 0.28.0"]}],"null_published_at":"2026-11-01T01:45:38.495Z","source_updated_at":"2026-10-06T00:02:07.000Z","source_disclosed_at":"2026-10-06T00:02:05.000Z","is_synthetic_fixture":false,"null_last_reviewed_at":null,"public_change_summary":"First indexed by Null; source disclosure date is shown separately.","null_first_observed_at":"2026-11-01T00:59:16.959Z"}],"next_cursor":null,"snapshot_revision":"42594"}