diff --git a/run.json b/run.json
index 303157b96..0ae0e11a6 100644
--- a/run.json
+++ b/run.json
@@ -196,14 +196,128 @@
     "in_place": false
   },
   "graph_source": "digraph Hello {\n    graph [goal=\"Say hello and demonstrate a basic Fabro workflow\"]\n    rankdir=LR\n\n    start [shape=Mdiamond, label=\"Start\"]\n    exit  [shape=Msquare, label=\"Exit\"]\n\n    greet [label=\"Greet\", prompt=\"Add a haiku to the README\"]\n\n    start -> greet -> exit\n}\n",
-  "start": null,
-  "status": {
-    "kind": "starting"
+  "start": {
+    "run_id": "01KQF5SZV94BZQG1JGK0YZ18KX",
+    "start_time": "2026-04-30T12:28:59.711476020Z",
+    "run_branch": "fabro/run/01KQF5SZV94BZQG1JGK0YZ18KX",
+    "base_sha": "272d2aa549c3b3ea31de59b24020a45dd41519d7"
   },
-  "status_updated_at": "2026-04-30T12:28:49.924890168Z",
+  "status": {
+    "kind": "running"
+  },
+  "status_updated_at": "2026-04-30T12:28:59.711556187Z",
   "pending_control": null,
-  "checkpoint": null,
-  "checkpoints": [],
+  "checkpoint": {
+    "timestamp": "2026-04-30T12:29:12.437340171Z",
+    "current_node": "greet",
+    "completed_nodes": [
+      "start",
+      "greet"
+    ],
+    "node_retries": {},
+    "context_values": {
+      "current_node": "greet",
+      "thread.start.current_node": "greet",
+      "internal.work_dir": "/workspace",
+      "internal.run_id": "01KQF5SZV94BZQG1JGK0YZ18KX",
+      "failure_signature": "",
+      "graph.goal": "Say hello and demonstrate a basic Fabro workflow",
+      "internal.fidelity": "compact",
+      "internal.retry_count.start": 0,
+      "response.greet": "Done! I added a haiku just below the tagline:\n\n```\nAgents write the code\nGraphs keep them on the right path\nYou review, then ship\n```\n\nIt captures the core Fabro philosophy in 5-7-5: AI agents do the work, workflow graphs provide the structure, and humans stay in control at the verification step.",
+      "last_stage": "greet",
+      "internal.node_visit_count": 1,
+      "failure_class": "",
+      "internal.retry_count.greet": 0,
+      "graph.rankdir": "LR",
+      "outcome": "succeeded",
+      "last_response": "Done! I added a haiku just below the tagline:\n\n```\nAgents write the code\nGraphs keep them on the right path\nYou review, then ship\n```\n\nIt captures the core Fabro philosophy in 5-7-5: AI agents do the ",
+      "internal.thread_id": "start"
+    },
+    "node_outcomes": {
+      "greet": {
+        "status": "succeeded",
+        "context_updates": {
+          "last_response": "Done! I added a haiku just below the tagline:\n\n```\nAgents write the code\nGraphs keep them on the right path\nYou review, then ship\n```\n\nIt captures the core Fabro philosophy in 5-7-5: AI agents do the ",
+          "response.greet": "Done! I added a haiku just below the tagline:\n\n```\nAgents write the code\nGraphs keep them on the right path\nYou review, then ship\n```\n\nIt captures the core Fabro philosophy in 5-7-5: AI agents do the work, workflow graphs provide the structure, and humans stay in control at the verification step.",
+          "last_stage": "greet"
+        },
+        "notes": "Stage completed: greet",
+        "usage": {
+          "input": {
+            "usage": {
+              "model": {
+                "provider": "anthropic",
+                "model_id": "claude-sonnet-4-6"
+              },
+              "tokens": {
+                "input_tokens": 4753,
+                "output_tokens": 408,
+                "reasoning_tokens": 0,
+                "cache_read_tokens": 23307,
+                "cache_write_tokens": 11962
+              }
+            },
+            "facts": {
+              "provider": "anthropic",
+              "cache_write_5m_tokens": 11962,
+              "cache_write_1h_tokens": 0
+            }
+          },
+          "total_usd_micros": 72228
+        },
+        "files_touched": [
+          "/workspace/README.md"
+        ]
+      },
+      "start": {
+        "status": "succeeded",
+        "usage": null
+      }
+    },
+    "next_node_id": "exit",
+    "node_visits": {
+      "start": 1,
+      "greet": 1
+    }
+  },
+  "checkpoints": [
+    [
+      20,
+      {
+        "timestamp": "2026-04-30T12:29:01.982564049Z",
+        "current_node": "start",
+        "completed_nodes": [
+          "start"
+        ],
+        "node_retries": {},
+        "context_values": {
+          "failure_signature": "",
+          "failure_class": "",
+          "outcome": "succeeded",
+          "graph.rankdir": "LR",
+          "internal.node_visit_count": 1,
+          "current_node": "start",
+          "internal.work_dir": "/workspace",
+          "internal.thread_id": null,
+          "internal.fidelity": "compact",
+          "internal.retry_count.start": 0,
+          "internal.run_id": "01KQF5SZV94BZQG1JGK0YZ18KX",
+          "graph.goal": "Say hello and demonstrate a basic Fabro workflow"
+        },
+        "node_outcomes": {
+          "start": {
+            "status": "succeeded",
+            "usage": null
+          }
+        },
+        "next_node_id": "greet",
+        "node_visits": {
+          "start": 1
+        }
+      }
+    ]
+  ],
   "conclusion": null,
   "retro": null,
   "retro_prompt": null,
@@ -220,5 +334,39 @@
   "pull_request": null,
   "superseded_by": null,
   "pending_interviews": {},
-  "nodes": {}
+  "nodes": {
+    "start@1": {
+      "prompt": null,
+      "response": null,
+      "status": {
+        "status": "succeeded",
+        "notes": null,
+        "failure_reason": null,
+        "timestamp": "2026-04-30T12:29:01.982511174Z"
+      },
+      "provider_used": null,
+      "diff": null,
+      "script_invocation": null,
+      "script_timing": null,
+      "parallel_results": null,
+      "stdout": null,
+      "stderr": null
+    },
+    "greet@1": {
+      "prompt": null,
+      "response": null,
+      "status": null,
+      "provider_used": {
+        "mode": "agent",
+        "provider": "anthropic",
+        "model": "claude-sonnet-4-6"
+      },
+      "diff": null,
+      "script_invocation": null,
+      "script_timing": null,
+      "parallel_results": null,
+      "stdout": null,
+      "stderr": null
+    }
+  }
 }
\ No newline at end of file
diff --git a/stages/greet@1/prompt.md b/stages/greet@1/prompt.md
new file mode 100644
index 000000000..e57c810c7
--- /dev/null
+++ b/stages/greet@1/prompt.md
@@ -0,0 +1,4 @@
+Goal: Say hello and demonstrate a basic Fabro workflow
+
+
+Add a haiku to the README
\ No newline at end of file
diff --git a/stages/greet@1/provider_used.json b/stages/greet@1/provider_used.json
new file mode 100644
index 000000000..d0418b4c6
--- /dev/null
+++ b/stages/greet@1/provider_used.json
@@ -0,0 +1,5 @@
+{
+  "mode": "agent",
+  "provider": "anthropic",
+  "model": "claude-sonnet-4-6"
+}
\ No newline at end of file
diff --git a/stages/greet@1/response.md b/stages/greet@1/response.md
new file mode 100644
index 000000000..5f371b11b
--- /dev/null
+++ b/stages/greet@1/response.md
@@ -0,0 +1,9 @@
+Done! I added a haiku just below the tagline:
+
+```
+Agents write the code
+Graphs keep them on the right path
+You review, then ship
+```
+
+It captures the core Fabro philosophy in 5-7-5: AI agents do the work, workflow graphs provide the structure, and humans stay in control at the verification step.
\ No newline at end of file
diff --git a/stages/start@1/status.json b/stages/start@1/status.json
new file mode 100644
index 000000000..a623721be
--- /dev/null
+++ b/stages/start@1/status.json
@@ -0,0 +1,6 @@
+{
+  "status": "succeeded",
+  "notes": null,
+  "failure_reason": null,
+  "timestamp": "2026-04-30T12:29:01.982511174Z"
+}
\ No newline at end of file