diff --git a/docs/receipts/external-pattern-research-map-20260702.json b/docs/receipts/external-pattern-research-map-20260702.json
new file mode 100644
index 0000000..ea90db5
--- /dev/null
+++ b/docs/receipts/external-pattern-research-map-20260702.json
@@ -0,0 +1,68 @@
+{
+  "schema": "fable.research_map.receipt.v1",
+  "createdAt": "2026-07-02T00:00:00.000Z",
+  "taskLabel": "external-pattern-research-map",
+  "task": "Map ECC, scankii, Miessler/Fable prompts, Test-don't-verify, and factory-memory correction into Fable engineering constraints",
+  "constraints": [
+    "external repositories and links are advisory until locally verified",
+    "do not vendor, copy, or execute external repositories from this research map",
+    "rewrite useful patterns locally behind receipts and tests",
+    "deploy is out of scope for this mapping",
+    "guarded 8099/deploy via git-proxy is the only canonical deploy route",
+    "8098/deploy-webhook remains legacy/internal/stale for fable-agent trust",
+    "live receipt > committed receipt > memory summary"
+  ],
+  "mappings": [
+    {
+      "name": "ECC",
+      "source": "https://github.com/affaan-m/ECC",
+      "localEvidence": "src/fable5/external-project-review.ts#createSkillsRepoReviewReceipt",
+      "patterns": ["agent harness optimization", "skills and memory organization", "research-first development posture"],
+      "adaptation": "learn-only; rewrite any useful harness ideas locally with Fable receipts as authority",
+      "blocked": ["vendoring", "external execution", "treating optimization claims as local proof"]
+    },
+    {
+      "name": "scankii",
+      "source": "https://github.com/ashp15205/scankii",
+      "localEvidence": "src/fable5/external-project-review.ts#createSkillsRepoReviewReceipt",
+      "patterns": ["AI-workflow SAST scanner ideas", "prompt-injection detection", "malicious tool payload review"],
+      "adaptation": "advisory rules only; rewrite useful checks locally with repro receipts/tests",
+      "blocked": ["vendoring", "unsandboxed scanner execution", "security claims without local reproduction"]
+    },
+    {
+      "name": "Miessler PAI and Fable prompt posture",
+      "source": "https://danielmiessler.com/",
+      "localEvidence": ["src/upgrades/pai-adapter.ts", "src/upgrades/fable5-prompt-engine.ts"],
+      "patterns": ["TELOS mapping", "harness-not-model compounding", "citation-aware prompt posture", "separate verifier instead of self-verification"],
+      "adaptation": "keep prompts as bounded operating guidance; receipts and verification remain authority",
+      "blocked": ["prompt-only deploy authority", "uncited factual claims from search", "self-verification as proof"]
+    },
+    {
+      "name": "Test-don't-verify",
+      "source": "TEMP/research/test-dont-verify/index.txt",
+      "localEvidence": "TEMP/research/test-dont-verify/index.txt",
+      "patterns": ["testing complements formal verification", "autoformalization expands the trusted computing base", "random testing/falsification finds counterexamples", "verification-guided development compares verified reference and production implementation"],
+      "adaptation": "pair proofs/specs with falsification, differential tests, and production-runtime measurements",
+      "blocked": ["claiming proof assistants replace tests", "trusting autoformalized specs without review", "performance claims without runtime evidence"]
+    },
+    {
+      "name": "factory-memory correction",
+      "source": "conversation memory summary quarantine",
+      "localEvidence": "src/fable5/stale-memory-quarantine.ts#createFactoryMemoryQuarantineReceipt",
+      "patterns": ["quarantine stale memory", "canonical trust path is receipt-backed", "8098 deploy memory is dangerous stale state"],
+      "adaptation": "prefer live/committed receipts over memory; require explicit approval and fresh gate receipts before any deploy",
+      "blocked": ["8098/deploy-webhook as trusted deploy path", "memory-only deploy authority", "autonomous deploy without fresh gate receipt"]
+    }
+  ],
+  "skipped": [
+    "external repo vendoring",
+    "external repo execution",
+    "deploy changes",
+    "dashboard changes",
+    "new generator or CLI",
+    "new tests without code parity need"
+  ],
+  "decision": "ready",
+  "reasons": [],
+  "deployAttempted": false
+}