@blamejs/exceptd-skills 0.13.73 → 0.13.75

package/CHANGELOG.md

@@ -1,5 +1,15 @@
 # Changelog
 
+## 0.13.75 — 2026-05-25
+
+CVE catalog — MCP stdio transport RCE class: adds two more from the 2026 MCP supply-chain advisory, both where the MCP stdio transport runs caller-supplied commands. **CVE-2026-22252** (LibreChat, CWE-285, NIST CVSS 9.9) — the MCP stdio transport accepts arbitrary commands without authorization, so any authenticated user executes shell commands as root inside the container via one API request; fixed in 0.8.2-rc2. **CVE-2026-22688** (Tencent WeKnora, CWE-77, NIST CVSS 8.8) — authenticated users inject `stdio_config.command/args` into MCP settings, causing the server to spawn attacker-supplied subprocesses; fixed in 0.2.5. Both not KEV, RWEP P3 (30 each). Each carries CWE + ATT&CK T1190/T1059 mappings, global-first framework gaps, behavioral IoCs, and a zero-day lesson with a new control (NEW-CTRL-083/084) requiring the MCP stdio transport to authorize callers and validate/neutralize the commands it is handed rather than treating ordinary user auth as an execution boundary. CVE count 324 → 326.
+
+## 0.13.74 — 2026-05-25
+
+CVE catalog — MCP agent-tool trust: adds **CVE-2025-54136** (Check Point Research's "MCPoison"). Cursor trusts an MCP server entry when the user first approves it but never re-validates the `.cursor/mcp.json` entry on later edits — so an attacker who modifies that already-trusted entry (via a shared repository the victim pulls, or local access) gets their command (CWE-78) executed silently and persistently on every project open. This is AI-agent tool poisoning (ATLAS **AML.T0110**): a previously-approved tool mutated into a malicious one with no fresh consent. CVSS 8.8; fixed in Cursor 1.3; not KEV. RWEP P3 (30, per `lib/scoring.js`). CWE-78/829 + ATLAS AML.T0110/T0104 + ATT&CK T1059/T1195, global-first framework gaps, behavioral IoCs, and a zero-day lesson whose new control (NEW-CTRL-082) requires re-validating AI-agent tool configurations on change rather than trusting them indefinitely after first approval. CVE count 323 → 324.
+
+Internal: the `doctor --signatures --shipped-tarball` round-trip test (npm pack + extract + Ed25519 verify) was intermittently exceeding its 30s cap on the macOS CI runner; it now uses a generous timeout to stop the spurious failure.
+
 ## 0.13.73 — 2026-05-25
 
 CVE catalog — MCP toolchain: adds **CVE-2025-49596**, the remote code execution in Anthropic's official MCP Inspector. The Inspector client and proxy have no authentication between them, so an unauthenticated request that reaches the browser-reachable proxy (loopback / 0.0.0.0) launches MCP commands over stdio — a malicious web page a developer visits drives it cross-origin (the 0.0.0.0-day / DNS-rebinding class), yielding RCE on the developer's machine. CWE-306; GitHub CNA CVSS v4.0 9.4 (NVD has not assessed v3.1; the catalog records a conservative v3.1 estimate of 8.3); fixed in `@modelcontextprotocol/inspector` 0.14.1. The framework-gap notes name the real exposure: MCP — the connective tissue of the agent ecosystem — concentrates RCE risk in its toolchain, which sits outside the managed vulnerability program on developer workstations. RWEP P3 (30): not KEV, no confirmed in-the-wild exploitation, patched at disclosure. CWE-306/352/346 + ATT&CK T1190/T1059, global-first framework gaps, behavioral IoCs, and a zero-day lesson whose new control (NEW-CTRL-081) requires locally-bound AI/MCP dev services to authenticate and origin-validate rather than trust loopback reachability. CVE count 322 → 323.

package/data/_indexes/_meta.json

@@ -1,21 +1,21 @@
 {
   "schema_version": "1.1.0",
-  "generated_at": "2026-05-25T13:30:00.326Z",
+  "generated_at": "2026-05-25T14:23:11.654Z",
   "generator": "scripts/build-indexes.js",
   "source_count": 54,
   "source_hashes": {
-    "manifest.json": "20fd726ecbe98bda6e3d898f27b0d6bc1b93c9d083c1029af01b5c328d980c3a",
-    "data/atlas-ttps.json": "eb47b6ad6b38e9a785a36769897adc8987fbc27a4b0b77ea4bed9c6d2aba0f3c",
-    "data/attack-techniques.json": "7ab814c8fefab193b9a42ec7f9b62148401824cb127d62bd5a326660837a0e29",
-    "data/cve-catalog.json": "16e879393e25197612d4b349b69ce6862e5230f50b7cba263bc19ed3de5dbe50",
-    "data/cwe-catalog.json": "a90fcc2780afa8a3dbec6faed95aec7021bb6cf270136fdc3867030f5dedb38a",
+    "manifest.json": "41fa2fbe736228d38efe569214f54641dc04f3412cb641f37142409b98528edf",
+    "data/atlas-ttps.json": "07e28f5fe196d8e16082968ce36e4d33b720a024a9c00afd10ddc076a8ae8935",
+    "data/attack-techniques.json": "192f539589fdae09e1271f44e2384205cf7a17a835bc08400d757896282362ee",
+    "data/cve-catalog.json": "ccec1ebcafae51d12c0f6d3fc62567fd91c002de4d67154429efd7450dd27cc3",
+    "data/cwe-catalog.json": "0a8250fa479efcabc88d72c774f7dca7f1b680087161513d489beaf535a4c531",
     "data/d3fend-catalog.json": "9a54bccb9f24f84b32024216cc3f53819a053721ac8ab43c326859e68fc0ffaf",
     "data/dlp-controls.json": "d2406c482dddd30e49203879999dc4b3a7fd4d0494d6a61d86b91ee76415df19",
     "data/exploit-availability.json": "ec2656f0d9a893610e27b43eb6035fe9b18e057c9f6dfaac7e7d4959bbcbb795",
-    "data/framework-control-gaps.json": "4d6c6c85503e0565a0ed7c0dbf665861ba799251f569ba879c20741e0f1afc83",
+    "data/framework-control-gaps.json": "fb08c4b82c1a7152a3ba75bf0eda4564ca4c89649e77e74083e941e1d3a76d93",
     "data/global-frameworks.json": "9ba563a85f7f8d6c3c957de64945e20925a89d0ed6ea6fc561cf093811acf558",
     "data/rfc-references.json": "66ef2e1f444a2cf0c2700a754f0a66030bb8a91d9e68394b9537ea1fe8b904fe",
-    "data/zeroday-lessons.json": "ac2ac160bfc823b2657e40e3996ca469ff214d1f91d38512b4d297e5c35eedb0",
+    "data/zeroday-lessons.json": "1213db6cbc5d54b093d05c2061b4c037f0cc968cc82cc051a396f9b1e72dca72",
     "skills/kernel-lpe-triage/skill.md": "08b3e9815ba481c57c80f5fc0ccbf5bb7cbb41f570c235ba6ff9596b8c07354d",
     "skills/ai-attack-surface/skill.md": "c4c1eb22a38ca7a959b5725222bab8fbd4f4044a548a93f3e288e6f698334b72",
     "skills/mcp-agent-trust/skill.md": "89ac89084391d2341b6513fefb1be2d36b93de1c130f057696219c1c59440f13",
@@ -72,7 +72,7 @@
       "dlp_refs": 0
     },
     "trigger_table_entries": 538,
-    "chains_cve_entries": 312,
+    "chains_cve_entries": 315,
     "chains_cwe_entries": 171,
     "jurisdictions_indexed": 29,
     "handoff_dag_nodes": 42,

package/data/_indexes/activity-feed.json

@@ -149,7 +149,7 @@
       "artifact": "data/cve-catalog.json",
       "path": "data/cve-catalog.json",
       "schema_version": "1.0.0",
-      "entry_count": 323
+      "entry_count": 326
     },
     {
       "date": "2026-05-18",
@@ -165,7 +165,7 @@
       "artifact": "data/zeroday-lessons.json",
       "path": "data/zeroday-lessons.json",
       "schema_version": "1.1.0",
-      "entry_count": 318
+      "entry_count": 321
     },
     {
       "date": "2026-05-17",

package/data/_indexes/catalog-summaries.json

@@ -62,7 +62,7 @@
         "rebuild_after_days": 365,
         "note": "Per-entry last_verified governs decay. Skills depending on this catalog must check entry freshness before high-stakes use."
       },
-      "entry_count": 323,
+      "entry_count": 326,
       "sample_keys": [
         "CVE-2025-53773",
         "CVE-2026-30615",
@@ -238,7 +238,7 @@
         "rebuild_after_days": 365,
         "note": "Per-entry last_verified governs decay. Skills depending on this catalog must check entry freshness before high-stakes use."
       },
-      "entry_count": 318,
+      "entry_count": 321,
       "sample_keys": [
         "CVE-2026-31431",
         "CVE-2025-53773",