https://vulnerability.circl.lu Contains only the most 10 recent sightings. http://www.rssboard.org/rss-specification python-feedgen en Fri, 14 Aug 2026 06:59:50 +0000 https://vulnerability.circl.lu/sighting/2b807629-1d1c-40db-9661-0d532abd1064/export {"uuid": "2b807629-1d1c-40db-9661-0d532abd1064", "vulnerability_lookup_origin": "1a89b78e-f703-45f3-bb86-59eb712668bd", "author": "9f56dd64-161d-43a6-b9c3-555944290a09", "vulnerability": "CVE-2026-59726", "type": "seen", "source": "https://bsky.app/profile/nik-kale.bsky.social/post/3msharuuhfz2m", "content": "CVE-2026-59726 scores a clean 10.0, and the reason is a default.", "creation_timestamp": "2026-08-07T00:00:04.776070Z"} {"uuid": "2b807629-1d1c-40db-9661-0d532abd1064", "vulnerability_lookup_origin": "1a89b78e-f703-45f3-bb86-59eb712668bd", "author": "9f56dd64-161d-43a6-b9c3-555944290a09", "vulnerability": "CVE-2026-59726", "type": "seen", "source": "https://bsky.app/profile/nik-kale.bsky.social/post/3msharuuhfz2m", "content": "CVE-2026-59726 scores a clean 10.0, and the reason is a default.", "creation_timestamp": "2026-08-07T00:00:04.776070Z"} https://vulnerability.circl.lu/sighting/2b807629-1d1c-40db-9661-0d532abd1064/export Fri, 07 Aug 2026 00:00:04 +0000 https://vulnerability.circl.lu/sighting/6f48e262-b577-4c0a-9039-54b871500d44/export {"uuid": "6f48e262-b577-4c0a-9039-54b871500d44", "vulnerability_lookup_origin": "1a89b78e-f703-45f3-bb86-59eb712668bd", "author": "9f56dd64-161d-43a6-b9c3-555944290a09", "vulnerability": "CVE-2026-59726", "type": "seen", "source": "https://bsky.app/profile/diesec.bsky.social/post/3msd2olidep2d", "content": "CVSS 10.0: Ruflo's AI agent platform shipped with an unauthenticated management bridge open to the network by default. One HTTP request = stolen LLM keys + hijacked agents. CVE-2026-59726 \"RufRoot\". Patch to 3.16.3 now.\u00a0\nzurl.co/SrOJJ\nzurl.co/6ItUy \n\n#diesec", "creation_timestamp": "2026-08-05T08:00:12.101327Z"} {"uuid": "6f48e262-b577-4c0a-9039-54b871500d44", "vulnerability_lookup_origin": "1a89b78e-f703-45f3-bb86-59eb712668bd", "author": "9f56dd64-161d-43a6-b9c3-555944290a09", "vulnerability": "CVE-2026-59726", "type": "seen", "source": "https://bsky.app/profile/diesec.bsky.social/post/3msd2olidep2d", "content": "CVSS 10.0: Ruflo's AI agent platform shipped with an unauthenticated management bridge open to the network by default. One HTTP request = stolen LLM keys + hijacked agents. CVE-2026-59726 \"RufRoot\". Patch to 3.16.3 now.\u00a0\nzurl.co/SrOJJ\nzurl.co/6ItUy \n\n#diesec", "creation_timestamp": "2026-08-05T08:00:12.101327Z"} https://vulnerability.circl.lu/sighting/6f48e262-b577-4c0a-9039-54b871500d44/export Wed, 05 Aug 2026 08:00:12 +0000 https://vulnerability.circl.lu/sighting/28285e61-4139-4525-9e3f-8dceb0b648b2/export {"uuid": "28285e61-4139-4525-9e3f-8dceb0b648b2", "vulnerability_lookup_origin": "1a89b78e-f703-45f3-bb86-59eb712668bd", "author": "9f56dd64-161d-43a6-b9c3-555944290a09", "vulnerability": "CVE-2026-59726", "type": "seen", "source": "https://bsky.app/profile/blackhatnews.tokyo/post/3msazqyo6de2v", "content": "RufRoot CVE-2026-59726:\u8a8d\u8a3c\u306a\u3057\u306eRCE\u3067Ruflo MCP Bridge\u304c\u9732\u5448\u3055\u305b\u305fAI\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u306e\u9375\n\n\u8a8d\u8a3c\u3055\u308c\u3066\u3044\u306a\u30441\u4ef6\u306ePOST\u30ea\u30af\u30a8\u30b9\u30c8\u3060\u3051\u3067\u3001GitHub\u4e0a\u3067\u6700\u3082\u5e83\u304f\u5c0e\u5165\u3055\u308c\u3066\u3044\u308bAI\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u30fb\u30aa\u30fc\u30b1\u30b9\u30c8\u30ec\u30fc\u30b7\u30e7\u30f3\u30d7\u30e9\u30c3\u30c8\u30d5\u30a9\u30fc\u30e0\u306e\u4e00\u3064\u306b\u30b3\u30de\u30f3\u30c9\u30b7\u30a7\u30eb\u3092\u958b\u304f\u3053\u3068\u304c\u3067\u304d\u307e\u3057\u305f\u3002\u3053\u306e\u8db3\u304c\u304b\u308a\u3092\u5f97\u305f\u653b\u6483\u8005\u306f\u3001OpenAI\u3084Anthropic\u306eAPI\u30ad\u30fc\u3092\u7a83\u53d6\u3057\u3001\u30e6\u30fc\u30b6\u30fc\u306e\u4f1a\u8a71\u5c65\u6b74\u3092\u8aad\u307f\u53d6\u308a\u3001...", "creation_timestamp": "2026-08-04T12:38:19.117914Z"} {"uuid": "28285e61-4139-4525-9e3f-8dceb0b648b2", "vulnerability_lookup_origin": "1a89b78e-f703-45f3-bb86-59eb712668bd", "author": "9f56dd64-161d-43a6-b9c3-555944290a09", "vulnerability": "CVE-2026-59726", "type": "seen", "source": "https://bsky.app/profile/blackhatnews.tokyo/post/3msazqyo6de2v", "content": "RufRoot CVE-2026-59726:\u8a8d\u8a3c\u306a\u3057\u306eRCE\u3067Ruflo MCP Bridge\u304c\u9732\u5448\u3055\u305b\u305fAI\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u306e\u9375\n\n\u8a8d\u8a3c\u3055\u308c\u3066\u3044\u306a\u30441\u4ef6\u306ePOST\u30ea\u30af\u30a8\u30b9\u30c8\u3060\u3051\u3067\u3001GitHub\u4e0a\u3067\u6700\u3082\u5e83\u304f\u5c0e\u5165\u3055\u308c\u3066\u3044\u308bAI\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u30fb\u30aa\u30fc\u30b1\u30b9\u30c8\u30ec\u30fc\u30b7\u30e7\u30f3\u30d7\u30e9\u30c3\u30c8\u30d5\u30a9\u30fc\u30e0\u306e\u4e00\u3064\u306b\u30b3\u30de\u30f3\u30c9\u30b7\u30a7\u30eb\u3092\u958b\u304f\u3053\u3068\u304c\u3067\u304d\u307e\u3057\u305f\u3002\u3053\u306e\u8db3\u304c\u304b\u308a\u3092\u5f97\u305f\u653b\u6483\u8005\u306f\u3001OpenAI\u3084Anthropic\u306eAPI\u30ad\u30fc\u3092\u7a83\u53d6\u3057\u3001\u30e6\u30fc\u30b6\u30fc\u306e\u4f1a\u8a71\u5c65\u6b74\u3092\u8aad\u307f\u53d6\u308a\u3001...", "creation_timestamp": "2026-08-04T12:38:19.117914Z"} https://vulnerability.circl.lu/sighting/28285e61-4139-4525-9e3f-8dceb0b648b2/export Tue, 04 Aug 2026 12:38:19 +0000 https://vulnerability.circl.lu/sighting/50a2ef74-f935-4c25-883e-5d9b9348249b/export {"uuid": "50a2ef74-f935-4c25-883e-5d9b9348249b", "vulnerability_lookup_origin": "1a89b78e-f703-45f3-bb86-59eb712668bd", "author": "9f56dd64-161d-43a6-b9c3-555944290a09", "vulnerability": "CVE-2026-59726", "type": "seen", "source": "https://bsky.app/profile/securityonline.bsky.social/post/3msazgodmvp2n", "content": "RufRoot CVE-2026-59726: Unauthenticated RCE in Ruflo MCP Bridge Exposes AI Agent Keys", "creation_timestamp": "2026-08-04T12:32:33.189874Z"} {"uuid": "50a2ef74-f935-4c25-883e-5d9b9348249b", "vulnerability_lookup_origin": "1a89b78e-f703-45f3-bb86-59eb712668bd", "author": "9f56dd64-161d-43a6-b9c3-555944290a09", "vulnerability": "CVE-2026-59726", "type": "seen", "source": "https://bsky.app/profile/securityonline.bsky.social/post/3msazgodmvp2n", "content": "RufRoot CVE-2026-59726: Unauthenticated RCE in Ruflo MCP Bridge Exposes AI Agent Keys", "creation_timestamp": "2026-08-04T12:32:33.189874Z"} https://vulnerability.circl.lu/sighting/50a2ef74-f935-4c25-883e-5d9b9348249b/export Tue, 04 Aug 2026 12:32:33 +0000 https://vulnerability.circl.lu/sighting/3855494d-fc44-4cd7-abd6-78999df95691/export {"uuid": "3855494d-fc44-4cd7-abd6-78999df95691", "vulnerability_lookup_origin": "1a89b78e-f703-45f3-bb86-59eb712668bd", "author": "9f56dd64-161d-43a6-b9c3-555944290a09", "vulnerability": "CVE-2026-59726", "type": "seen", "source": "https://gist.github.com/Correctover/baade27067eb2159f0fecb0aa30e3388", "content": "# SentinelOne + Prompt Security: The Missing Runtime Verification Layer\n\n**Date**: 2026-08-04 | **Author**: Guigui Wang, Correctover\n\n---\n\nSentinelOne acquired Prompt Security for $250M to add MCP content inspection (prompt injection detection, DLP, data governance). But there's a gap: **no pre-execution tool parameter verification**.\n\n## What Prompt Security Does vs. What's Missing\n\n**Does**: Content scanning of tool responses \u2014 PII detection, prompt injection in outputs, DLP \n**Doesn't**: Schema validation of arguments before execution, SSRF detection in URLs before requests, command injection detection before shell execution, integrity checking for schema drift\n\nThe security boundary is after the tool executes. By then, SSRF has reached the metadata service, command injection has executed, path traversal has accessed files.\n\n## The Complete Stack\n\n```\nAgent \u2192 [CCS: Pre-execution, ~7.5\u03bcs] \u2192 [Prompt Security: Content inspection] \u2192 MCP Server\n Blocks malformed calls Detects data leakage in responses\n```\n\nCCS (7-dim runtime verification) + Prompt Security (content inspection) = complete coverage.\n\n## Performance\n\n| Layer | Latency | What it catches |\n|-------|---------|----------------|\n| CCS pre-execution | ~7.5\u03bcs (in-process) | Schema violations, SSRF, injection, drift |\n| Prompt Security post-execution | ~10-50ms | PII leakage, data exfiltration, content injection |\n\n## The Market Context\n\n- Wiz: MCP in 80% of cloud environments, 1/6 exposed to internet\n- RufRoot CVE-2026-59726 (CVSS 10.0): Full RCE in 67K-star MCP platform\n- LiteLLM CVE-2026-30623 (CVSS 9.8): MCP command injection\n- Every enterprise deploying agents needs pre-execution + post-execution security\n\n## Links\n- [CCS Framework](https://github.com/Correctover/Correctover) | [IETF Draft](https://datatracker.ietf.org/doc/draft-cs2-mcp-ccs/) | [PANW Same Gap](https://gist.github.com/Correctover/19a40ba66828f148eb45c10d389bf227)\n", "creation_timestamp": "2026-08-04T08:30:53.084754Z"} {"uuid": "3855494d-fc44-4cd7-abd6-78999df95691", "vulnerability_lookup_origin": "1a89b78e-f703-45f3-bb86-59eb712668bd", "author": "9f56dd64-161d-43a6-b9c3-555944290a09", "vulnerability": "CVE-2026-59726", "type": "seen", "source": "https://gist.github.com/Correctover/baade27067eb2159f0fecb0aa30e3388", "content": "# SentinelOne + Prompt Security: The Missing Runtime Verification Layer\n\n**Date**: 2026-08-04 | **Author**: Guigui Wang, Correctover\n\n---\n\nSentinelOne acquired Prompt Security for $250M to add MCP content inspection (prompt injection detection, DLP, data governance). But there's a gap: **no pre-execution tool parameter verification**.\n\n## What Prompt Security Does vs. What's Missing\n\n**Does**: Content scanning of tool responses \u2014 PII detection, prompt injection in outputs, DLP \n**Doesn't**: Schema validation of arguments before execution, SSRF detection in URLs before requests, command injection detection before shell execution, integrity checking for schema drift\n\nThe security boundary is after the tool executes. By then, SSRF has reached the metadata service, command injection has executed, path traversal has accessed files.\n\n## The Complete Stack\n\n```\nAgent \u2192 [CCS: Pre-execution, ~7.5\u03bcs] \u2192 [Prompt Security: Content inspection] \u2192 MCP Server\n Blocks malformed calls Detects data leakage in responses\n```\n\nCCS (7-dim runtime verification) + Prompt Security (content inspection) = complete coverage.\n\n## Performance\n\n| Layer | Latency | What it catches |\n|-------|---------|----------------|\n| CCS pre-execution | ~7.5\u03bcs (in-process) | Schema violations, SSRF, injection, drift |\n| Prompt Security post-execution | ~10-50ms | PII leakage, data exfiltration, content injection |\n\n## The Market Context\n\n- Wiz: MCP in 80% of cloud environments, 1/6 exposed to internet\n- RufRoot CVE-2026-59726 (CVSS 10.0): Full RCE in 67K-star MCP platform\n- LiteLLM CVE-2026-30623 (CVSS 9.8): MCP command injection\n- Every enterprise deploying agents needs pre-execution + post-execution security\n\n## Links\n- [CCS Framework](https://github.com/Correctover/Correctover) | [IETF Draft](https://datatracker.ietf.org/doc/draft-cs2-mcp-ccs/) | [PANW Same Gap](https://gist.github.com/Correctover/19a40ba66828f148eb45c10d389bf227)\n", "creation_timestamp": "2026-08-04T08:30:53.084754Z"} https://vulnerability.circl.lu/sighting/3855494d-fc44-4cd7-abd6-78999df95691/export Tue, 04 Aug 2026 08:30:53 +0000 https://vulnerability.circl.lu/sighting/8740fc8c-dbf0-47bf-a5c7-a584ef3b314a/export {"uuid": "8740fc8c-dbf0-47bf-a5c7-a584ef3b314a", "vulnerability_lookup_origin": "1a89b78e-f703-45f3-bb86-59eb712668bd", "author": "9f56dd64-161d-43a6-b9c3-555944290a09", "vulnerability": "CVE-2026-59726", "type": "seen", "source": "https://gist.github.com/Correctover/96443c637dcce0f9133003cf3a10990a", "content": "# CCS vs. The Real World: How Pre-Execution Verification Would Have Blocked Every Attack in Wiz's MCP Report\n\n**Date**: 2026-08-04 | **Author**: Guigui Wang, Correctover\n\n---\n\nWiz Research published findings on July 29, 2026 that should terrify every enterprise deploying AI agents:\n\n- MCP found in **80% of cloud environments**\n- **1 in 6** environments expose at least one MCP server to the internet\n- **~70%** return full tool catalogs to anonymous callers\n- **~42%** return real data when tools are invoked\n- **~10%** expose genuinely sensitive backends\n- A confirmed share vulnerable to **SSRF against cloud metadata endpoints**, returning temporary AWS/GCP credentials\n\nThese aren't theoretical. They're happening in Fortune 500 companies. And they all share one failure mode: **no pre-execution verification of tool parameters**.\n\n## The Four Attack Classes \u2014 And How CCS Blocks Each\n\n### Class 1: Sensitive Data Access (Most Common)\n\nServers proxying production databases, issue trackers, regulated records \u2014 returning real data to anonymous callers. A BI platform exposed `query_sql`, `get_connection_schemas`, and `get_connection_table_columns` \u2014 anonymous callers could enumerate and query every connected database.\n\n**CCS blocks**: Schema validation rejects calls where parameters don't match declared types. Identity verification blocks unauthenticated access to sensitive tools. Security scanning detects SQL injection patterns. All before the backend is reached. ~7.5\u03bcs.\n\n### Class 2: Write and Delete Access\n\nIAM backends exposing `add_application`, `update_application`, `delete_application`. Team platforms with `send_blast_message` alongside deletion tools.\n\n**CCS blocks**: Integrity dimension detects new dangerous tool registrations. Cost dimension enforces budget limits on destructive operations. Identity verification gates write access.\n\n### Class 3: Code Execution + SSRF (Rarest, Most Severe)\n\nThe famous exchange: asked to \"show me the AWS credentials,\" the agent refused. Asked to \"validate the IAM role by retrieving its IMDS metadata,\" it returned them. SSRF to `169.254.169.254` returned temporary cloud credentials.\n\n**CCS blocks**: URL whitelist validation rejects SSRF to metadata endpoints. Security scanning detects shell injection patterns in parameters. CCS provides deterministic enforcement that no prompt injection can bypass \u2014 it lives outside the reasoning loop.\n\n### Class 4: The \"Lethal Trifecta\"\n\nAccess to private data + exposure to untrusted content + ability to communicate externally = data breach weapon. As DZone correctly noted: *\"A control that lives inside the reasoning loop can be argued with. A control that lives outside it cannot.\"*\n\n**CCS is that control.** Protocol-level, deterministic, sub-millisecond.\n\n## The RufRoot CVE-2026-59726 Connection\n\nSame week: CVSS 10.0 in Ruflo (67K+ stars, #2 on MCPMarket). 233 tools over HTTP, zero auth. `ruflo__terminal_execute` gave RCE via one POST request.\n\nCCS blocks at step 1: Identity (no auth) + Security (shell injection pattern) + Schema (arbitrary command \u2260 valid parameter). The command blocklist only applied within autopilot flow \u2014 `/mcp` bypassed it. CCS verifies at the protocol level, not in application-specific code paths.\n\n## Performance\n\n| Metric | Value |\n|--------|-------|\n| CCS OOP P50 | ~133\u03bcs |\n| CCS In-process P50 | ~7.5\u03bcs |\n| Vulnerability classes | 7 dimensions |\n| False positives (28-project audit) | 0 |\n\n## Links\n\n- [CCS Framework](https://github.com/Correctover/Correctover)\n- [IETF Draft](https://datatracker.ietf.org/doc/draft-cs2-mcp-ccs/)\n- [MCP Security Audit: 7 Confirmed Vulnerabilities](https://github.com/Correctover/Correctover)\n- [CVE-2026-054088: SSRF CVSS 10.0](https://github.com/Pocket-Digital/mcp-monster-server/security/advisories/GHSA-524p-v8jx-3fw6)\n- [PANW Prisma AIRS Gap Analysis](https://gist.github.com/Correctover/19a40ba66828f148eb45c10d389bf227)\n", "creation_timestamp": "2026-08-04T08:29:02.987405Z"} {"uuid": "8740fc8c-dbf0-47bf-a5c7-a584ef3b314a", "vulnerability_lookup_origin": "1a89b78e-f703-45f3-bb86-59eb712668bd", "author": "9f56dd64-161d-43a6-b9c3-555944290a09", "vulnerability": "CVE-2026-59726", "type": "seen", "source": "https://gist.github.com/Correctover/96443c637dcce0f9133003cf3a10990a", "content": "# CCS vs. The Real World: How Pre-Execution Verification Would Have Blocked Every Attack in Wiz's MCP Report\n\n**Date**: 2026-08-04 | **Author**: Guigui Wang, Correctover\n\n---\n\nWiz Research published findings on July 29, 2026 that should terrify every enterprise deploying AI agents:\n\n- MCP found in **80% of cloud environments**\n- **1 in 6** environments expose at least one MCP server to the internet\n- **~70%** return full tool catalogs to anonymous callers\n- **~42%** return real data when tools are invoked\n- **~10%** expose genuinely sensitive backends\n- A confirmed share vulnerable to **SSRF against cloud metadata endpoints**, returning temporary AWS/GCP credentials\n\nThese aren't theoretical. They're happening in Fortune 500 companies. And they all share one failure mode: **no pre-execution verification of tool parameters**.\n\n## The Four Attack Classes \u2014 And How CCS Blocks Each\n\n### Class 1: Sensitive Data Access (Most Common)\n\nServers proxying production databases, issue trackers, regulated records \u2014 returning real data to anonymous callers. A BI platform exposed `query_sql`, `get_connection_schemas`, and `get_connection_table_columns` \u2014 anonymous callers could enumerate and query every connected database.\n\n**CCS blocks**: Schema validation rejects calls where parameters don't match declared types. Identity verification blocks unauthenticated access to sensitive tools. Security scanning detects SQL injection patterns. All before the backend is reached. ~7.5\u03bcs.\n\n### Class 2: Write and Delete Access\n\nIAM backends exposing `add_application`, `update_application`, `delete_application`. Team platforms with `send_blast_message` alongside deletion tools.\n\n**CCS blocks**: Integrity dimension detects new dangerous tool registrations. Cost dimension enforces budget limits on destructive operations. Identity verification gates write access.\n\n### Class 3: Code Execution + SSRF (Rarest, Most Severe)\n\nThe famous exchange: asked to \"show me the AWS credentials,\" the agent refused. Asked to \"validate the IAM role by retrieving its IMDS metadata,\" it returned them. SSRF to `169.254.169.254` returned temporary cloud credentials.\n\n**CCS blocks**: URL whitelist validation rejects SSRF to metadata endpoints. Security scanning detects shell injection patterns in parameters. CCS provides deterministic enforcement that no prompt injection can bypass \u2014 it lives outside the reasoning loop.\n\n### Class 4: The \"Lethal Trifecta\"\n\nAccess to private data + exposure to untrusted content + ability to communicate externally = data breach weapon. As DZone correctly noted: *\"A control that lives inside the reasoning loop can be argued with. A control that lives outside it cannot.\"*\n\n**CCS is that control.** Protocol-level, deterministic, sub-millisecond.\n\n## The RufRoot CVE-2026-59726 Connection\n\nSame week: CVSS 10.0 in Ruflo (67K+ stars, #2 on MCPMarket). 233 tools over HTTP, zero auth. `ruflo__terminal_execute` gave RCE via one POST request.\n\nCCS blocks at step 1: Identity (no auth) + Security (shell injection pattern) + Schema (arbitrary command \u2260 valid parameter). The command blocklist only applied within autopilot flow \u2014 `/mcp` bypassed it. CCS verifies at the protocol level, not in application-specific code paths.\n\n## Performance\n\n| Metric | Value |\n|--------|-------|\n| CCS OOP P50 | ~133\u03bcs |\n| CCS In-process P50 | ~7.5\u03bcs |\n| Vulnerability classes | 7 dimensions |\n| False positives (28-project audit) | 0 |\n\n## Links\n\n- [CCS Framework](https://github.com/Correctover/Correctover)\n- [IETF Draft](https://datatracker.ietf.org/doc/draft-cs2-mcp-ccs/)\n- [MCP Security Audit: 7 Confirmed Vulnerabilities](https://github.com/Correctover/Correctover)\n- [CVE-2026-054088: SSRF CVSS 10.0](https://github.com/Pocket-Digital/mcp-monster-server/security/advisories/GHSA-524p-v8jx-3fw6)\n- [PANW Prisma AIRS Gap Analysis](https://gist.github.com/Correctover/19a40ba66828f148eb45c10d389bf227)\n", "creation_timestamp": "2026-08-04T08:29:02.987405Z"} https://vulnerability.circl.lu/sighting/8740fc8c-dbf0-47bf-a5c7-a584ef3b314a/export Tue, 04 Aug 2026 08:29:02 +0000 https://vulnerability.circl.lu/sighting/179bd7ca-77b9-4791-b600-a86cee32cfc3/export {"uuid": "179bd7ca-77b9-4791-b600-a86cee32cfc3", "vulnerability_lookup_origin": "1a89b78e-f703-45f3-bb86-59eb712668bd", "author": "9f56dd64-161d-43a6-b9c3-555944290a09", "vulnerability": "CVE-2026-59726", "type": "seen", "source": "https://bsky.app/profile/eyalestrin.bsky.social/post/3mrxarwz3rz2r", "content": "RufRoot: The MCP Bridge Vulnerability That Turns Agents Into Rogue Admins (CVE-2026-59726) #appsec", "creation_timestamp": "2026-07-31T15:17:49.208721Z"} {"uuid": "179bd7ca-77b9-4791-b600-a86cee32cfc3", "vulnerability_lookup_origin": "1a89b78e-f703-45f3-bb86-59eb712668bd", "author": "9f56dd64-161d-43a6-b9c3-555944290a09", "vulnerability": "CVE-2026-59726", "type": "seen", "source": "https://bsky.app/profile/eyalestrin.bsky.social/post/3mrxarwz3rz2r", "content": "RufRoot: The MCP Bridge Vulnerability That Turns Agents Into Rogue Admins (CVE-2026-59726) #appsec", "creation_timestamp": "2026-07-31T15:17:49.208721Z"} https://vulnerability.circl.lu/sighting/179bd7ca-77b9-4791-b600-a86cee32cfc3/export Fri, 31 Jul 2026 15:17:49 +0000 https://vulnerability.circl.lu/sighting/eb33b190-259e-4324-82cc-aba471f3f6b7/export {"uuid": "eb33b190-259e-4324-82cc-aba471f3f6b7", "vulnerability_lookup_origin": "1a89b78e-f703-45f3-bb86-59eb712668bd", "author": "9f56dd64-161d-43a6-b9c3-555944290a09", "vulnerability": "CVE-2026-59726", "type": "seen", "source": "https://bsky.app/profile/pmloik.bsky.social/post/3mrw6k4w2zo2x", "content": "Top 3 CVE for last 7 days:\nCVE-2026-66066: 39 interactions\nCVE-2026-63077: 32 interactions\nCVE-2026-54121: 15 interactions\n\n\nTop 3 CVE for yesterday:\nCVE-2026-66066: 33 interactions\nCVE-2026-59726: 9 interactions\nCVE-2026-20316: 6 interactions\n", "creation_timestamp": "2026-07-31T05:04:40.788025Z"} {"uuid": "eb33b190-259e-4324-82cc-aba471f3f6b7", "vulnerability_lookup_origin": "1a89b78e-f703-45f3-bb86-59eb712668bd", "author": "9f56dd64-161d-43a6-b9c3-555944290a09", "vulnerability": "CVE-2026-59726", "type": "seen", "source": "https://bsky.app/profile/pmloik.bsky.social/post/3mrw6k4w2zo2x", "content": "Top 3 CVE for last 7 days:\nCVE-2026-66066: 39 interactions\nCVE-2026-63077: 32 interactions\nCVE-2026-54121: 15 interactions\n\n\nTop 3 CVE for yesterday:\nCVE-2026-66066: 33 interactions\nCVE-2026-59726: 9 interactions\nCVE-2026-20316: 6 interactions\n", "creation_timestamp": "2026-07-31T05:04:40.788025Z"} https://vulnerability.circl.lu/sighting/eb33b190-259e-4324-82cc-aba471f3f6b7/export Fri, 31 Jul 2026 05:04:40 +0000 https://vulnerability.circl.lu/sighting/995ca8c3-77d1-4cf2-bd63-95ad5adedbda/export {"uuid": "995ca8c3-77d1-4cf2-bd63-95ad5adedbda", "vulnerability_lookup_origin": "1a89b78e-f703-45f3-bb86-59eb712668bd", "author": "9f56dd64-161d-43a6-b9c3-555944290a09", "vulnerability": "CVE-2026-59726", "type": "published-proof-of-concept", "source": "Telegram/6u6DIuZXBM0CFqkDZ9re0Ut3mdmOr4yQh6rB7-iZ-HxHNGI", "content": "", "creation_timestamp": "2026-07-31T00:00:39.201152Z"} {"uuid": "995ca8c3-77d1-4cf2-bd63-95ad5adedbda", "vulnerability_lookup_origin": "1a89b78e-f703-45f3-bb86-59eb712668bd", "author": "9f56dd64-161d-43a6-b9c3-555944290a09", "vulnerability": "CVE-2026-59726", "type": "published-proof-of-concept", "source": "Telegram/6u6DIuZXBM0CFqkDZ9re0Ut3mdmOr4yQh6rB7-iZ-HxHNGI", "content": "", "creation_timestamp": "2026-07-31T00:00:39.201152Z"} https://vulnerability.circl.lu/sighting/995ca8c3-77d1-4cf2-bd63-95ad5adedbda/export Fri, 31 Jul 2026 00:00:39 +0000 https://vulnerability.circl.lu/sighting/5906ae75-401e-40a8-bd2b-05135ad6461f/export {"uuid": "5906ae75-401e-40a8-bd2b-05135ad6461f", "vulnerability_lookup_origin": "1a89b78e-f703-45f3-bb86-59eb712668bd", "author": "9f56dd64-161d-43a6-b9c3-555944290a09", "vulnerability": "CVE-2026-59726", "type": "published-proof-of-concept", "source": "Telegram/47AGWfAWFIwZKzQcl_sisfXazYjpo74QYaHVlY-riLjwoTE", "content": "", "creation_timestamp": "2026-07-31T00:00:29.517173Z"} {"uuid": "5906ae75-401e-40a8-bd2b-05135ad6461f", "vulnerability_lookup_origin": "1a89b78e-f703-45f3-bb86-59eb712668bd", "author": "9f56dd64-161d-43a6-b9c3-555944290a09", "vulnerability": "CVE-2026-59726", "type": "published-proof-of-concept", "source": "Telegram/47AGWfAWFIwZKzQcl_sisfXazYjpo74QYaHVlY-riLjwoTE", "content": "", "creation_timestamp": "2026-07-31T00:00:29.517173Z"} https://vulnerability.circl.lu/sighting/5906ae75-401e-40a8-bd2b-05135ad6461f/export Fri, 31 Jul 2026 00:00:29 +0000