diff --git a/README.md b/README.md
index d310f64..a0d0e76 100644
--- a/README.md
+++ b/README.md
@@ -18,6 +18,7 @@ Platform tooling, **access documentation**, and Jira planning for the Platypus (
 
 - [PLAT board](https://catalystsoftware.atlassian.net/jira/software/c/projects/PLAT/summary)
 - [Access docs](access/) — endpoints, agents, MCP map (scaffold; synthesis ticket pending)
+- [Galva program](docs/research/galva-program.md) — modernization research via `platform-galva` (E10 / [PLAT-96](https://catalystsoftware.atlassian.net/browse/PLAT-96))
 - [Security policy](docs/security.md) — no secrets in git
 
 ## Status
diff --git a/docs/how-we-work.md b/docs/how-we-work.md
index 10741da..03b65ab 100644
--- a/docs/how-we-work.md
+++ b/docs/how-we-work.md
@@ -46,7 +46,7 @@ cp config/config.yaml.example ~/.config/platform-tools/config.yaml
 
 Override the path with `PLATFORM_TOOLS_CONFIG` if needed. **Never** commit `~/.config/platform-tools/config.yaml` or copy its contents into PRs, Jira tickets, or agent context.
 
-The same convention applies to other Platypus repos (`platform-ikg`, `platform-castleguard`, `platform-zaha`): each uses `~/.config/<repo_name>/config.yaml` for local secrets and endpoints.
+The same convention applies to other Platypus repos (`platform-ikg`, `platform-castleguard`, `platform-zaha`, `platform-galva`): each uses `~/.config/<repo_name>/config.yaml` for local secrets and endpoints.
 
 ## Privacy
 
@@ -82,6 +82,21 @@ Access docs and Jira tickets in this repo must respect **personal privacy**, **c
 
 Scaffold-only content is expected until the synthesis ticket ([PLAT-92](https://catalystsoftware.atlassian.net/browse/PLAT-92)) is completed via privacy-safe PRs.
 
+## Galva — modernization research (E10)
+
+**Galva** ([PLAT-96](https://catalystsoftware.atlassian.net/browse/PLAT-96)) is the PLAT program for researching how to modernize, migrate, and refactor older codebases — without executing production changes in the research phase. The future implementation repo is **`totango/platform-galva`** (same naming pattern as `platform-ikg`, `platform-castleguard`, `platform-zaha`).
+
+| Concept | Guidance |
+|---------|----------|
+| **Repo** | `totango/platform-galva` — research tooling and intake automation (not yet scaffolded) |
+| **Intake** | Facts in (topology, deps, versions) → strategy out (savings, effort, risk, blast radius) |
+| **Sizing** | Small galvas (one program) vs big galvas (service groups) — see [galva-program.md](research/galva-program.md) |
+| **Languages** | Java, Python, Go, etc. — derive ASTs per ecosystem; no single-language assumption |
+| **Platform context** | Use IKG (E09), FinOps (E04), access docs — not personal machine paths |
+| **Handoff** | Approved plans become delivery stories under E01/E02/… — Galva does not merge infra changes |
+
+Example inaugural galva: [catalyst-ingest → shared Temporal](research/galva-catalyst-ingest.md) ([PLAT-98](https://catalystsoftware.atlassian.net/browse/PLAT-98)).
+
 ## Deployment and merge process
 
 `main` is **protected** on Platypus repos. Treat merges as a small deployment: traceable ticket, human review, and CI validation.
@@ -101,7 +116,7 @@ Apply or update protection on a repo:
 .github/scripts/setup-branch-protection.sh totango/platform-tools
 ```
 
-Use the same script (with a different `owner/repo` argument) when bootstrapping `platform-ikg`, `platform-castleguard`, `platform-zaha`, and other Platypus repos.
+Use the same script (with a different `owner/repo` argument) when bootstrapping `platform-ikg`, `platform-castleguard`, `platform-zaha`, `platform-galva`, and other Platypus repos.
 
 ### PR title (Jira-linked)
 
diff --git a/docs/jira-space-overview.md b/docs/jira-space-overview.md
index 8132eab..5f7b710 100644
--- a/docs/jira-space-overview.md
+++ b/docs/jira-space-overview.md
@@ -22,7 +22,7 @@ Epic (initiative — quarter / multi-month)
 
 **Rule:** Acceptance criteria on **Story**; definition of done on **Tasks**.
 
-## Epic map (E01–E09)
+## Epic map (E01–E10)
 
 | Epic | Name |
 |------|------|
@@ -35,8 +35,11 @@ Epic (initiative — quarter / multi-month)
 | E07 | Agent research & validation |
 | E08 | Agent platform & lifecycle |
 | E09 | Infra Knowledge Graph (IKG) & Zaha UI |
+| E10 | Galva — modernization research & migration planning ([PLAT-96](https://catalystsoftware.atlassian.net/browse/PLAT-96)) |
 
-Traceability labels: `plat-ref:E01` … `plat-ref:E09`, `plat-ref:100` for logical IDs.
+Traceability labels: `plat-ref:E01` … `plat-ref:E10`, `plat-ref:100` for logical IDs.
+
+**Galva (E10):** research/planning for modernizing older codebases — future repo `totango/platform-galva`; see [docs/research/galva-program.md](research/galva-program.md).
 
 ## Tooling in this repo
 
diff --git a/docs/research/galva-catalyst-ingest.md b/docs/research/galva-catalyst-ingest.md
new file mode 100644
index 0000000..d473ef6
--- /dev/null
+++ b/docs/research/galva-catalyst-ingest.md
@@ -0,0 +1,110 @@
+---
+title: "Galva example — catalyst-ingest migrate to shared Temporal"
+tags: [eng-information, galva, catalyst-ingest, temporal, migration]
+last_updated: "2026-08-28"
+status: scaffold
+audience: [engineers, agents]
+gaps:
+  - "Current Temporal version and namespace layout not verified"
+  - "PostgreSQL sizing, HA model, and migration tooling unknown"
+  - "Shared Temporal target cluster/version not selected"
+  - "Consumer blast radius not inventoried"
+sources_to_synthesize:
+  - "Approved service repo (catalyst-ingest) — via PR, no personal paths"
+  - "Platform IKG when available (E09)"
+  - "FinOps signals for dedicated vs shared Temporal cost (E04)"
+jira: "PLAT-98"
+---
+
+# Galva: catalyst-ingest → shared Temporal
+
+**Jira:** [PLAT-98](https://catalystsoftware.atlassian.net/browse/PLAT-98) (Story under [PLAT-96](https://catalystsoftware.atlassian.net/browse/PLAT-96) / E10)  
+**Size:** Small-to-medium galva (single primary workload; shared infra dependency)  
+**Program:** [galva-program.md](./galva-program.md)
+
+## Goal
+
+Research whether **catalyst-ingest** should migrate from its **dedicated Temporal deployment** (possibly older version) to a **shared Temporal platform**, including PostgreSQL requirements and EKS placement constraints.
+
+This document is a **research scaffold** — facts marked `TBD` must be verified before any delivery epic is opened.
+
+## Known facts (intake)
+
+| Field | Value | Verified |
+|-------|-------|----------|
+| Workload | `catalyst-ingest` | Partial — name only |
+| EKS placement | Dedicated cluster, **same AWS account and VPC** as shared services | `TBD` |
+| Workflow engine | Temporal (version `TBD` — possibly behind current shared standard) | `TBD` |
+| Data store | PostgreSQL required | `TBD` |
+| Languages | `TBD` (likely Java and/or Go — confirm from repo) | `TBD` |
+
+## Research questions
+
+### 1. Current state
+
+| Question | Notes |
+|----------|-------|
+| Which EKS cluster and namespace? | Same-VPC constraint affects network path to shared Temporal |
+| Temporal version, namespace, and task queue layout? | Version skew drives SDK and server compatibility matrix |
+| PostgreSQL: RDS vs in-cluster? Region? HA? | Co-migration vs split-brain risk |
+| Workflow/activity inventory | Count, criticality, long-running vs cron |
+| Who consumes ingest outputs? | Blast radius for cutover |
+
+### 2. Target state options
+
+| Option | Pros | Cons |
+|--------|------|------|
+| **A. Shared Temporal cluster, new namespace** | Lower ops cost; centralized upgrades | Noisy-neighbor; RBAC/isolation design needed |
+| **B. Shared Temporal + version upgrade in place** | Single migration event | Higher risk; longer freeze window |
+| **C. Stay dedicated; upgrade only** | Minimal blast radius | Does not capture shared-platform savings |
+
+**Deliverable:** Recommended option with rollback path.
+
+### 3. Cost and efficiency
+
+| Line item | Dedicated (today) | Shared (hypothesis) |
+|-----------|-------------------|---------------------|
+| Temporal compute / storage | `TBD` | `TBD` |
+| PostgreSQL | `TBD` | `TBD` |
+| Engineering toil (upgrades, on-call) | Qualitative | Qualitative |
+
+Coordinate with **FinOps (E04)** for utilization data — do not paste billing exports into git.
+
+### 4. Effort, risk, blast radius
+
+| Dimension | Initial hypothesis |
+|-----------|-------------------|
+| **Effort** | `TBD` — depends on SDK version gap and DB migration approach |
+| **Risk** | Workflow nondeterminism on SDK upgrade; DB schema migration downtime |
+| **Blast radius** | All workflows and downstream consumers of ingest artifacts |
+| **Dependencies** | Shared Temporal readiness; VPC/security group rules; DB access (see [PLAT-95](https://catalystsoftware.atlassian.net/browse/PLAT-95)) |
+
+### 5. Suggested migration phases (draft)
+
+1. **Discover** — AST/repo analysis + IKG topology; verify intake table above
+2. **Design** — target namespace, RBAC, DB strategy, compatibility matrix
+3. **Pilot** — non-prod namespace on shared Temporal; shadow traffic or canary workflows
+4. **Cutover** — phased queue migration with rollback triggers
+5. **Decommission** — retire dedicated Temporal resources after soak period
+
+## Language / AST focus
+
+When analyzing the catalyst-ingest repo:
+
+- Locate Temporal **workflow** and **activity** registrations
+- Map **PostgreSQL** access (ORM, migrations, connection pools)
+- Identify **deployment config** (Helm/K8s, env vars — placeholders only in docs)
+- Flag **version-sensitive** APIs (Temporal SDK changelog)
+
+## Out of scope
+
+- Executing production migration (opens under delivery epic after galva sign-off)
+- Committing kubeconfig, connection strings, or customer data
+- Bulk-creating Jira tasks without human review
+
+## Acceptance criteria (PLAT-98)
+
+- [ ] Intake table verified or gaps explicitly documented
+- [ ] Target-state recommendation with cost/effort/risk summary
+- [ ] Up to 5 follow-up delivery tasks **proposed** (not auto-created)
+- [ ] All examples obfuscated; no secrets in git
diff --git a/docs/research/galva-program.md b/docs/research/galva-program.md
new file mode 100644
index 0000000..12b56d0
--- /dev/null
+++ b/docs/research/galva-program.md
@@ -0,0 +1,139 @@
+---
+title: "Galva — modernization research & migration planning"
+tags: [eng-information, galva, modernization, migration]
+last_updated: "2026-08-28"
+status: draft
+audience: [engineers, agents]
+gaps:
+  - "totango/platform-galva repo does not exist yet — program brief lives in platform-tools until scaffolded"
+  - "Per-language AST tooling choices (Java, Python, Go) not finalized"
+  - "Intake automation and IKG integration paths TBD"
+sources_to_synthesize:
+  - "argocd-tele: workspace/docs/jira-space-organization.md (E03 RFW, E04 FinOps, E09 IKG)"
+  - "platform-tools/docs/research/galva-catalyst-ingest.md (example inaugural galva)"
+jira: "PLAT-96"
+---
+
+# Galva program brief
+
+**Jira Epic:** [PLAT-96](https://catalystsoftware.atlassian.net/browse/PLAT-96) (E10 — Galva)  
+**Future repo:** `totango/platform-galva` (not created in this pass)  
+**Board:** [PLAT](https://catalystsoftware.atlassian.net/jira/software/c/projects/PLAT/summary)
+
+## What is Galva?
+
+**Galva** is a platform program for **research and planning** on how to modernize, improve cost efficiency, migrate, and refactor older codebases. Galva does **not** execute production changes — it produces strategies, estimates, and risk analysis for humans to approve and deliver under separate epics.
+
+```
+Facts in (topology, deps, versions, constraints)
+        │
+        ▼
+┌───────────────────┐
+│  Galva research   │  AST / static analysis per language
+│  + platform ctx   │  deployment patterns, VPC/EKS limits
+└─────────┬─────────┘
+          │
+          ▼
+Strategy out: savings · effort · risk · blast radius · phases
+```
+
+## Sizing: big vs small galvas
+
+| Size | Scope | Example | Typical output |
+|------|-------|---------|----------------|
+| **Small galva** | Single program, one service, narrow infra move | Move one worker to shared Temporal namespace | 1–2 week research; focused migration plan |
+| **Big galva** | Service group, cross-cutting refactor, multi-cluster | Consolidate ingest pipeline onto shared platform services | Multi-sprint research; phased program with checkpoints |
+
+Use **effort**, **risk**, and **blast radius** together — a small code change with wide consumer impact may still be a **big galva**.
+
+## Intake template (facts in → strategy out)
+
+Every galva starts with a structured intake. Fill what is known; mark unknowns explicitly.
+
+| Field | Description | Example (placeholder) |
+|-------|-------------|-------------------------|
+| **Workload name** | Service or repo identifier | `catalyst-ingest` |
+| **Languages / runtimes** | Primary stacks | Java 11, Go 1.21, Python 3.9 |
+| **Deployment** | Cluster, namespace, account (no secrets) | Dedicated EKS in `<AWS_ACCOUNT>`, same VPC as shared services |
+| **Dependencies** | Data stores, queues, external APIs | PostgreSQL, Temporal `<VERSION>`, S3 |
+| **Consumers** | Who depends on this workload | Downstream `<SERVICE_A>`, batch jobs |
+| **Constraints** | VPC, region, compliance, uptime | Same-VPC only; no cross-region DB today |
+| **Cost signals** | Known waste or duplication | Dedicated Temporal cluster vs shared |
+| **Goal** | What “better” looks like | Migrate to shared Temporal; upgrade workflow SDK |
+
+**Output sections** (every galva deliverable):
+
+1. Current-state architecture (verified or marked `TBD`)
+2. Target-state options (≥2 where feasible)
+3. Cost / efficiency comparison (qualitative OK for v1)
+4. Effort estimate (T-shirt + dependencies)
+5. Risk register + blast radius
+6. Recommended phases + rollback strategy
+7. Proposed follow-up delivery tickets (human review — no bulk create)
+
+## Language-agnostic analysis
+
+Galva must work across **Java, Python, Go**, and other stacks without assuming one toolchain.
+
+| Ecosystem | Analysis approach (research) | Derives |
+|-----------|------------------------------|---------|
+| **Java** | AST via parser (e.g. JavaParser) or build-tool graph (Maven/Gradle) | Service boundaries, Temporal workflow/activity usage, JDBC deps |
+| **Python** | `ast` module, import graph, optional type hints | Worker entrypoints, ORM usage, async patterns |
+| **Go** | `go/ast`, module graph, static call graph | Temporal SDK usage, gRPC clients, config loading |
+| **Infra** | Terraform/K8s manifests, Helm values (from approved repos) | Cluster placement, IAM, network paths |
+
+Galva should **compose** language-specific extractors behind a common intake schema rather than one monolithic analyzer.
+
+## Platform context Galva needs
+
+Galva research quality depends on deployment and platform knowledge:
+
+| Source | Use |
+|--------|-----|
+| **IKG (E09)** | Service topology, AWS/K8s/TF relationships |
+| **Castleguard** | Boundary and security findings |
+| **FinOps (E04)** | Cost baselines and savings hypotheses |
+| **RFW (E03)** | Proactive debt that migrations should not repeat |
+| **Access docs** | How engineers reach clusters, DBs, shared services |
+| **PLAT-95** | Teleport / multi-cluster / DB access patterns |
+
+## Relationship to other PLAT epics
+
+```mermaid
+flowchart LR
+  E10[PLAT-E10 Galva]
+  E03[E03 RFW]
+  E04[E04 FinOps]
+  E09[E09 IKG]
+  DEL[Delivery epics E01/E02/...]
+
+  E09 --> E10
+  E04 --> E10
+  E03 --> E10
+  E10 -->|approved plans| DEL
+```
+
+- **Galva (E10)** — research and migration **planning**
+- **RFW (E03)** — prevent repeat pain; galva findings often spawn RFW stories
+- **FinOps (E04)** — savings validation; Zaha may feed cost signals into intake
+- **Delivery epics** — execute approved migrations after galva sign-off
+
+## Privacy and content rules
+
+- No customer names, tenant IDs, or production hostnames in git or Jira bodies
+- No secrets, tokens, or personal machine paths
+- Use `~/.config/<repo_name>/config.yaml` for local tooling references only
+- Obfuscate examples: `<CUSTOMER>`, `<TENANT_ID>`, `<CLUSTER_NAME>`
+
+## Stories under E10
+
+| Key | Summary |
+|-----|---------|
+| [PLAT-97](https://catalystsoftware.atlassian.net/browse/PLAT-97) | platform-galva — intake model and multi-language research scaffold |
+| [PLAT-98](https://catalystsoftware.atlassian.net/browse/PLAT-98) | Galva: catalyst-ingest migrate to shared Temporal (example inaugural galva) |
+
+## Next steps
+
+1. Review and approve [PR for galva docs](https://github.com/totango/platform-tools/pulls) (branch `ab/galva-epic`)
+2. Scaffold `totango/platform-galva` repo when ready (separate ticket)
+3. Execute PLAT-98 research brief — [galva-catalyst-ingest.md](./galva-catalyst-ingest.md)
diff --git a/jira/plans/examples/galva-epic-ticket-draft.md b/jira/plans/examples/galva-epic-ticket-draft.md
new file mode 100644
index 0000000..c2baffb
--- /dev/null
+++ b/jira/plans/examples/galva-epic-ticket-draft.md
@@ -0,0 +1,21 @@
+# Galva epic — PLAT-96
+
+**Key:** [PLAT-96](https://catalystsoftware.atlassian.net/browse/PLAT-96)  
+**Type:** Epic (E10)  
+**Labels:** `plat-ref:epic-E10`, `galva`, `modernization`, `eng-information`, `migration`
+
+## Summary
+
+Galva — modernization research & migration planning
+
+## Child stories
+
+| Key | Summary |
+|-----|---------|
+| [PLAT-97](https://catalystsoftware.atlassian.net/browse/PLAT-97) | platform-galva — intake model and multi-language research scaffold |
+| [PLAT-98](https://catalystsoftware.atlassian.net/browse/PLAT-98) | Galva: catalyst-ingest migrate to shared Temporal |
+
+## Canonical docs
+
+- `docs/research/galva-program.md`
+- `docs/research/galva-catalyst-ingest.md`
diff --git a/jira/plans/examples/snapshot-fixture.json b/jira/plans/examples/snapshot-fixture.json
index e2074dd..99cdd65 100644
--- a/jira/plans/examples/snapshot-fixture.json
+++ b/jira/plans/examples/snapshot-fixture.json
@@ -17,6 +17,11 @@
       "ref": "E09",
       "summary": "Infra Knowledge Graph (IKG) & Zaha UI",
       "labels": ["plat-ref:E09", "ikg"]
+    },
+    {
+      "ref": "E10",
+      "summary": "Galva — modernization research & migration planning",
+      "labels": ["plat-ref:E10", "galva", "modernization"]
     }
   ],
   "issues": [
diff --git a/jira/rules/dependency-rules.yaml b/jira/rules/dependency-rules.yaml
index ad22b95..e9d2358 100644
--- a/jira/rules/dependency-rules.yaml
+++ b/jira/rules/dependency-rules.yaml
@@ -46,3 +46,27 @@ rules:
     to: { epic_ref: E01 }
     link: relates_to
     reason: "RFW stories often harden E01/E02 delivery — link with relates to in Jira"
+
+  - id: galva-consumes-ikg
+    from: { epic_ref: E10 }
+    to: { epic_ref: E09 }
+    link: consumes
+    reason: "Galva research uses IKG for service topology and deployment context"
+
+  - id: galva-finops-signals
+    from: { epic_ref: E04 }
+    to: { epic_ref: E10 }
+    link: feeds
+    reason: "FinOps cost signals inform Galva intake and savings hypotheses"
+
+  - id: galva-rfw-output
+    from: { epic_ref: E10 }
+    to: { epic_ref: E03 }
+    link: relates_to
+    reason: "Approved Galva plans often spawn RFW stories to prevent repeat migration pain"
+
+  - id: galva-delivery-handoff
+    from: { epic_ref: E10 }
+    to: { epic_ref: E01 }
+    link: handoff
+    reason: "Galva produces plans; delivery epics execute approved migrations"
diff --git a/jira/rules/epic-preservation.yaml b/jira/rules/epic-preservation.yaml
index a3a1740..b3493af 100644
--- a/jira/rules/epic-preservation.yaml
+++ b/jira/rules/epic-preservation.yaml
@@ -38,6 +38,9 @@ epics:
   - ref: E09
     summary_contains: "Infra Knowledge Graph"
     labels: [plat-ref:epic-E09, ikg]
+  - ref: E10
+    summary_contains: "Galva"
+    labels: [plat-ref:epic-E10, galva, modernization, migration]
 
 on_conflict:
   action: skip_and_log