diff --git a/.github/workflows/pricing-staleness.yml b/.github/workflows/pricing-staleness.yml
new file mode 100644
index 00000000..32fef033
--- /dev/null
+++ b/.github/workflows/pricing-staleness.yml
@@ -0,0 +1,31 @@
+# Scheduled freshness gate for the migration-to-aws pricing caches.
+#
+# The estimate phases price from vendored caches whose freshness contracts
+# (_meta.last_updated + staleness_days, or a "Last updated" line with a
+# documented 30-day window) were previously unenforced — caches quietly crossed
+# their own thresholds and every estimate silently degraded to fallback
+# accuracy. `mise run build` reports staleness warn-only (a stale cache must
+# not fail unrelated PRs); this weekly job runs the same tool with --strict so
+# staleness becomes a visible failure someone owns, prompting a data-refresh PR
+# (see #131 for the pattern).
+name: Pricing cache staleness
+on:
+  schedule:
+    - cron: "23 14 * * 1" # weekly, Monday
+  workflow_dispatch: {}
+permissions:
+  actions: none
+  contents: none
+jobs:
+  staleness:
+    permissions:
+      contents: read
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@08c6903cd8c0fde910a37f88322edcfb5dd907a8 # v5.0.0
+      - uses: actions/setup-node@49933ea5288caeca8642d1e84afbd3f7d6820020 # v4.4.0
+        with:
+          node-version: 24
+      # Zero-dep TS tool (Node 24 native type-stripping) — no install step.
+      - name: Check pricing cache freshness (strict)
+        run: node migrate/plugins/migration-to-aws/tools/pricing-staleness.ts --strict
diff --git a/migrate/plugins/migration-to-aws/skills/heroku-to-aws/references/shared/heroku-pricing-cache.md b/migrate/plugins/migration-to-aws/skills/heroku-to-aws/references/shared/heroku-pricing-cache.md
index e6c0bb40..d66eb6a0 100644
--- a/migrate/plugins/migration-to-aws/skills/heroku-to-aws/references/shared/heroku-pricing-cache.md
+++ b/migrate/plugins/migration-to-aws/skills/heroku-to-aws/references/shared/heroku-pricing-cache.md
@@ -1,6 +1,6 @@
 # Heroku Pricing Cache
 
-**Last updated:** 2026-06-15
+**Last updated:** 2026-07-19 (dyno rates re-verified against devcenter.heroku.com/articles/dyno-types — Eco $5 flat, Basic $7, Standard-1X $25, Standard-2X $50, Performance-M $250, Performance-L/L-RAM $500 all unchanged; no rate changes this refresh)
 **Source:** https://elements.heroku.com/addons/heroku-postgresql, https://elements.heroku.com/addons/heroku-redis, https://elements.heroku.com/addons/heroku-kafka, https://devcenter.heroku.com/articles/dyno-sizes
 **Currency:** USD
 **Accuracy:** ±5% for dynos (published flat rates); ±10% for data services (Elements "Max of" pricing, actual may vary by usage pattern)
diff --git a/migrate/plugins/migration-to-aws/skills/heroku-to-aws/references/vendored/pricing/aws-infra-pricing.json b/migrate/plugins/migration-to-aws/skills/heroku-to-aws/references/vendored/pricing/aws-infra-pricing.json
index b297f909..1f483c78 100644
--- a/migrate/plugins/migration-to-aws/skills/heroku-to-aws/references/vendored/pricing/aws-infra-pricing.json
+++ b/migrate/plugins/migration-to-aws/skills/heroku-to-aws/references/vendored/pricing/aws-infra-pricing.json
@@ -7,7 +7,8 @@
     "accuracy": "\u00b15-10% infrastructure",
     "hours_per_month": 730,
     "staleness_days": 30,
-    "staleness_note": "If more than staleness_days past last_updated, infrastructure rates remain reliable; set pricing_source=cached_stale and note it. AI-model rates (not used for heroku->aws) may have drifted."
+    "staleness_note": "If more than staleness_days past last_updated, infrastructure rates remain reliable; set pricing_source=cached_stale and note it. AI-model rates (not used for heroku->aws) may have drifted.",
+    "last_verification": "2026-07-19 spot-check against published pricing pages: Fargate Linux/x86 us-east-1 ($0.000011244/vCPU-s = $0.04048/hr; $0.000001235/GB-s = the published $0.004445/hr GB rate) matches, and Heroku dyno flat rates match devcenter; no rate changes applied this refresh."
   },
   "_multi_az_convention": {
     "_comment": "THE TRAP: a design's multi_az:true means different things per service. Each service rate below carries multi_az_handling so the formula applies the RIGHT adjustment, never a blanket multiplier.",
diff --git a/migrate/plugins/migration-to-aws/skills/shared/pricing/aws-infra-pricing.json b/migrate/plugins/migration-to-aws/skills/shared/pricing/aws-infra-pricing.json
index b297f909..1f483c78 100644
--- a/migrate/plugins/migration-to-aws/skills/shared/pricing/aws-infra-pricing.json
+++ b/migrate/plugins/migration-to-aws/skills/shared/pricing/aws-infra-pricing.json
@@ -7,7 +7,8 @@
     "accuracy": "\u00b15-10% infrastructure",
     "hours_per_month": 730,
     "staleness_days": 30,
-    "staleness_note": "If more than staleness_days past last_updated, infrastructure rates remain reliable; set pricing_source=cached_stale and note it. AI-model rates (not used for heroku->aws) may have drifted."
+    "staleness_note": "If more than staleness_days past last_updated, infrastructure rates remain reliable; set pricing_source=cached_stale and note it. AI-model rates (not used for heroku->aws) may have drifted.",
+    "last_verification": "2026-07-19 spot-check against published pricing pages: Fargate Linux/x86 us-east-1 ($0.000011244/vCPU-s = $0.04048/hr; $0.000001235/GB-s = the published $0.004445/hr GB rate) matches, and Heroku dyno flat rates match devcenter; no rate changes applied this refresh."
   },
   "_multi_az_convention": {
     "_comment": "THE TRAP: a design's multi_az:true means different things per service. Each service rate below carries multi_az_handling so the formula applies the RIGHT adjustment, never a blanket multiplier.",
diff --git a/migrate/plugins/migration-to-aws/skills/vercel-to-aws/references/vendored/pricing/aws-infra-pricing.json b/migrate/plugins/migration-to-aws/skills/vercel-to-aws/references/vendored/pricing/aws-infra-pricing.json
index b297f909..1f483c78 100644
--- a/migrate/plugins/migration-to-aws/skills/vercel-to-aws/references/vendored/pricing/aws-infra-pricing.json
+++ b/migrate/plugins/migration-to-aws/skills/vercel-to-aws/references/vendored/pricing/aws-infra-pricing.json
@@ -7,7 +7,8 @@
     "accuracy": "\u00b15-10% infrastructure",
     "hours_per_month": 730,
     "staleness_days": 30,
-    "staleness_note": "If more than staleness_days past last_updated, infrastructure rates remain reliable; set pricing_source=cached_stale and note it. AI-model rates (not used for heroku->aws) may have drifted."
+    "staleness_note": "If more than staleness_days past last_updated, infrastructure rates remain reliable; set pricing_source=cached_stale and note it. AI-model rates (not used for heroku->aws) may have drifted.",
+    "last_verification": "2026-07-19 spot-check against published pricing pages: Fargate Linux/x86 us-east-1 ($0.000011244/vCPU-s = $0.04048/hr; $0.000001235/GB-s = the published $0.004445/hr GB rate) matches, and Heroku dyno flat rates match devcenter; no rate changes applied this refresh."
   },
   "_multi_az_convention": {
     "_comment": "THE TRAP: a design's multi_az:true means different things per service. Each service rate below carries multi_az_handling so the formula applies the RIGHT adjustment, never a blanket multiplier.",