systeminit · stack72 · Mar 11, 2026 · Mar 11, 2026 · Mar 11, 2026 · Mar 11, 2026
diff --git a/.claude/skills/swamp-data/SKILL.md b/.claude/skills/swamp-data/SKILL.md
@@ -1,6 +1,6 @@
 ---
 name: swamp-data
-description: Manage model data lifecycle and garbage collection. Use when listing data, viewing versions, cleaning up old data, or understanding data retention. Triggers on "data", "swamp data", "model data", "data list", "data get", "data versions", "garbage collection", "gc", "data gc", "clean up data", "old data", "data retention", "data lifecycle", "version history", "data cleanup", "prune data", "expire data", "ephemeral data".
+description: List model data, view version history, delete expired versions, and run garbage collection. Use when listing data, viewing versions, cleaning up old data, or configuring data retention. Triggers on "swamp data", "model data", "data list", "data get", "data versions", "garbage collection", "gc", "data gc", "clean up data", "old data", "data retention", "data lifecycle", "version history", "data cleanup", "prune data", "expire data", "ephemeral data".
 ---
 
 # Swamp Data Skill
@@ -340,134 +340,15 @@ swamp data gc -f --json  # Skip confirmation prompt
 
 ## Accessing Data in Expressions
 
-Use CEL expressions to access model data in workflows and model inputs.
-
-**Note:** `model.<name>.resource.<spec>` requires the model to have previously
-produced data (a method was run that called `writeResource`). If no data exists
-yet, accessing `.resource` will fail with "No such key". Use
-`swamp data list <model-name>` to verify data exists.
-
-```yaml
-# Access latest resource data via dot notation
-value: ${{ model.my-model.resource.output.main.attributes.result }}
-
-# Access specific version
-value: ${{ data.version("my-model", "main", 2).attributes.result }}
-
-# Access file metadata
-path: ${{ model.my-model.file.content.primary.path }}
-size: ${{ model.my-model.file.content.primary.size }}
-
-# Lazy-load file contents
-body: ${{ file.contents("my-model", "content") }}
-```
-
-### Data Namespace Functions
-
-| Function                                     | Description                               |
-| -------------------------------------------- | ----------------------------------------- |
-| `data.version(modelName, dataName, version)` | Get specific version of data              |
-| `data.latest(modelName, dataName)`           | Get latest version of data                |
-| `data.listVersions(modelName, dataName)`     | Get array of available version numbers    |
-| `data.findByTag(tagKey, tagValue)`           | Find all data matching a tag              |
-| `data.findBySpec(modelName, specName)`       | Find all data from a specific output spec |
-
-**DataRecord structure** returned by these functions:
-
-```json
-{
-  "id": "uuid",
-  "name": "data-name",
-  "version": 3,
-  "createdAt": "2025-01-15T10:30:00Z",
-  "attributes": {/* data content */},
-  "tags": { "type": "resource" }
-}
-```
-
-**Example usage:**
-
-```yaml
-# Get specific version
-oldValue: ${{ data.version("my-model", "state", 2).attributes.value }}
-
-# Get latest
-current: ${{ data.latest("my-model", "output").attributes.result }}
-
-# List versions for conditional logic
-hasHistory: ${{ size(data.listVersions("my-model", "state")) > 1 }}
-
-# Find all resources across models
-allResources: ${{ data.findByTag("type", "resource") }}
-
-# Find data from a specific workflow
-workflowData: ${{ data.findByTag("workflow", "my-workflow") }}
-
-# Find all instances from a factory model's output spec
-subnets: ${{ data.findBySpec("my-scanner", "subnet") }}
-```
-
-**Key rules:**
-
-- `model.<name>.resource.<specName>.<instanceName>` — accesses the latest
-  version of a resource. Works both within a workflow run (in-memory updates)
-  and across workflow runs (persisted data).
-- `model.<name>.file.<specName>.<instanceName>` — accesses file metadata (path,
-  size, contentType). Same behavior as resource expressions.
-- `data.latest(modelName, dataName)` — reads persisted data snapshot taken at
-  workflow start.
-- Use `data.version()` function for specific versions
-- Use `data.findByTag()` to query across models
-- See the `swamp-workflow` skill's
-  [data-chaining reference](../swamp-workflow/references/data-chaining.md) for
-  detailed guidance on expression choice in workflows.
+CEL expressions access model data in workflows and model inputs. Functions,
+examples, and key rules are in
+[references/expressions.md](references/expressions.md).
 
 ## Data Ownership
 
-Data artifacts are owned by the model (definition) that created them. This
-ensures data integrity and prevents accidental overwrites.
-
-### Owner Definition
-
-Each data item tracks its owner through the `ownerDefinition` field:
-
-| Field            | Description                                  |
-| ---------------- | -------------------------------------------- |
-| `ownerType`      | `model-method`, `workflow-step`, or `manual` |
-| `ownerRef`       | Reference to the creating entity             |
-| `definitionHash` | Hash of the definition at creation time      |
-| `workflowId`     | Set when created during workflow execution   |
-| `workflowRunId`  | Specific run that created this data          |
-
-### Ownership Validation
-
-When a model method writes data:
-
-1. **New data**: Created with current model as owner
-2. **Existing data**: Validates `ownerDefinition.definitionHash` matches
-3. **Hash mismatch**: Write fails with ownership error
-
-This prevents scenarios where multiple models accidentally share data names.
-
-### Viewing Ownership
-
-Use `swamp data get` to see ownership information:
-
-```bash
-swamp data get my-model state --json
-```
-
-```json
-{
-  "name": "state",
-  "version": 3,
-  "ownerDefinition": {
-    "ownerType": "model-method",
-    "ownerRef": "my-model:create",
-    "definitionHash": "abc123..."
-  }
-}
-```
+Data is owned by the creating model — see
+[references/data-ownership.md](references/data-ownership.md) for owner fields,
+validation rules, and viewing ownership.
 
 ## Data Storage
 

diff --git a/.claude/skills/swamp-data/references/data-ownership.md b/.claude/skills/swamp-data/references/data-ownership.md
@@ -0,0 +1,46 @@
+# Data Ownership
+
+Data artifacts are owned by the model (definition) that created them. This
+ensures data integrity and prevents accidental overwrites.
+
+## Owner Definition
+
+Each data item tracks its owner through the `ownerDefinition` field:
+
+| Field            | Description                                  |
+| ---------------- | -------------------------------------------- |
+| `ownerType`      | `model-method`, `workflow-step`, or `manual` |
+| `ownerRef`       | Reference to the creating entity             |
+| `definitionHash` | Hash of the definition at creation time      |
+| `workflowId`     | Set when created during workflow execution   |
+| `workflowRunId`  | Specific run that created this data          |
+
+## Ownership Validation
+
+When a model method writes data:
+
+1. **New data**: Created with current model as owner
+2. **Existing data**: Validates `ownerDefinition.definitionHash` matches
+3. **Hash mismatch**: Write fails with ownership error
+
+This prevents scenarios where multiple models accidentally share data names.
+
+## Viewing Ownership
+
+Use `swamp data get` to see ownership information:
+
+```bash
+swamp data get my-model state --json
+```
+
+```json
+{
+  "name": "state",
+  "version": 3,
+  "ownerDefinition": {
+    "ownerType": "model-method",
+    "ownerRef": "my-model:create",
+    "definitionHash": "abc123..."
+  }
+}
+```
diff --git a/.claude/skills/swamp-data/references/expressions.md b/.claude/skills/swamp-data/references/expressions.md
@@ -0,0 +1,83 @@
+# Accessing Data in Expressions
+
+Use CEL expressions to access model data in workflows and model inputs.
+
+**Note:** `model.<name>.resource.<spec>` requires the model to have previously
+produced data (a method was run that called `writeResource`). If no data exists
+yet, accessing `.resource` will fail with "No such key". Use
+`swamp data list <model-name>` to verify data exists.
+
+```yaml
+# Access latest resource data via dot notation
+value: ${{ model.my-model.resource.output.main.attributes.result }}
+
+# Access specific version
+value: ${{ data.version("my-model", "main", 2).attributes.result }}
+
+# Access file metadata
+path: ${{ model.my-model.file.content.primary.path }}
+size: ${{ model.my-model.file.content.primary.size }}
+
+# Lazy-load file contents
+body: ${{ file.contents("my-model", "content") }}
+```
+
+## Data Namespace Functions
+
+| Function                                     | Description                               |
+| -------------------------------------------- | ----------------------------------------- |
+| `data.version(modelName, dataName, version)` | Get specific version of data              |
+| `data.latest(modelName, dataName)`           | Get latest version of data                |
+| `data.listVersions(modelName, dataName)`     | Get array of available version numbers    |
+| `data.findByTag(tagKey, tagValue)`           | Find all data matching a tag              |
+| `data.findBySpec(modelName, specName)`       | Find all data from a specific output spec |
+
+**DataRecord structure** returned by these functions:
+
+```json
+{
+  "id": "uuid",
+  "name": "data-name",
+  "version": 3,
+  "createdAt": "2025-01-15T10:30:00Z",
+  "attributes": {/* data content */},
+  "tags": { "type": "resource" }
+}
+```
+
+**Example usage:**
+
+```yaml
+# Get specific version
+oldValue: ${{ data.version("my-model", "state", 2).attributes.value }}
+
+# Get latest
+current: ${{ data.latest("my-model", "output").attributes.result }}
+
+# List versions for conditional logic
+hasHistory: ${{ size(data.listVersions("my-model", "state")) > 1 }}
+
+# Find all resources across models
+allResources: ${{ data.findByTag("type", "resource") }}
+
+# Find data from a specific workflow
+workflowData: ${{ data.findByTag("workflow", "my-workflow") }}
+
+# Find all instances from a factory model's output spec
+subnets: ${{ data.findBySpec("my-scanner", "subnet") }}
+```
+
+**Key rules:**
+
+- `model.<name>.resource.<specName>.<instanceName>` — accesses the latest
+  version of a resource. Works both within a workflow run (in-memory updates)
+  and across workflow runs (persisted data).
+- `model.<name>.file.<specName>.<instanceName>` — accesses file metadata (path,
+  size, contentType). Same behavior as resource expressions.
+- `data.latest(modelName, dataName)` — reads persisted data snapshot taken at
+  workflow start.
+- Use `data.version()` function for specific versions
+- Use `data.findByTag()` to query across models
+- See the `swamp-workflow` skill's
+  [data-chaining reference](../../swamp-workflow/references/data-chaining.md)
+  for detailed guidance on expression choice in workflows.
diff --git a/.claude/skills/swamp-extension-model/SKILL.md b/.claude/skills/swamp-extension-model/SKILL.md
@@ -1,6 +1,6 @@
 ---
 name: swamp-extension-model
-description: Create user-defined TypeScript models for swamp. Use when users want to extend swamp with custom model types, create automation models, or add new integrations. Triggers on "create model", "new model type", "custom model", "extension model", "user model", "typescript model", "extend swamp", "build integration", "zod schema", "model plugin", "deno model", "extensions/models", "model development", "implement model".
+description: Create user-defined TypeScript models for swamp — define Zod schemas, implement model interfaces, configure output specs. Use when users want to extend swamp with custom model types, create automation models, or add new integrations. Triggers on "create model", "new model type", "custom model", "extension model", "user model", "typescript model", "extend swamp", "build integration", "zod schema", "model plugin", "deno model", "extensions/models", "model development", "implement model".
 ---
 
 # Swamp Extension Model