# Cxas Dfcx Migration

> Migrate Dialogflow CX (DFCX) agents to CXAS (Customer Experience Agent Studio) agents. Use this skill when the user mentions DFCX migration, migrating agents, converting DFCX to CXAS, porting agents, agent migration, or post-migration optimization/consolidation. Four independently runnable scripts: migrate.py (1:1), stage_1.py (variable dedup + consolidation), stage_2.py (instruction state machines + tool mocks + lint + report), stage_3.py (rewires consolidated topology from source dep graph; only needed when stage_1 ran consolidation). State persists between scripts via <target>_ir.json so each can run / re-run / resume independently.

- Skill: `googlecloudplatform/cxas-dfcx-migration` (Agent Skill, multi-file: 10 files)
- Install (CLI): `npx skillmds@latest add googlecloudplatform/cxas-dfcx-migration`
- Raw SKILL.md: https://api.skillmd.com/api/skills/googlecloudplatform/cxas-dfcx-migration/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: AI & ML
- Author: GoogleCloudPlatform (https://skillmd.com/u/googlecloudplatform)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/googlecloudplatform/cxas-dfcx-migration

---


# DFCX to CXAS Migration

Four small scripts, one persistent IR bundle:

| Script | What it does | Runtime | Output |
|---|---|---|---|
| `migrate.py` | 1:1 conversion of every selected playbook/flow into the IR, and deploys **base resources only** (app, variables, tools). **Agent deployment is DEFERRED to `stage_1.py`** so large sources don't exceed the CXAS 100-agent cap — the compiled agents are saved in `<target>_ir.json`, not pushed. Pass `--no-consolidate` to push the full 1:1 agent set immediately (only safe below ~100 agents). | ~30 min for ~40 flows | `<target>_ir.json`, `<target>_migration_report.md`, `<target>_unit_tests.json` |
| `stage_1.py` | Loads the IR bundle, runs `CXASOptimizer.optimize_stage1` (variable dedup) and Gemini structural consolidation (N→M agent grouping). **This is the first agent push to CXAS** — only the consolidated (N→M) agents are deployed; the raw 1:1 originals are never pushed (`consolidate()` drops them and the pre-consolidation snapshot is used transiently for the integrity check only, never persisted). CXAS Version `0.0.2` (dedup) and `0.0.3` (consolidation). | ~15 min | Updated `<target>_ir.json`, `<target>_grouping.json` |
| `stage_2.py` | Loads the IR bundle, runs `CXASOptimizer.optimize_stage2` (instruction state machines + tool mocks). Pushes via update-pass deploys. CXAS Version `0.0.4`. Re-generates unit tests. Lints. Writes the audit report. | ~10 min | Updated `<target>_ir.json`, `<target>_optimization_report.md`, regenerated `<target>_unit_tests.json` |
| `stage_3.py` | **Only after Stage 1 consolidation.** Rewires the consolidated agents' parent → children topology by mapping the SOURCE DFCX dep graph onto the new groups (rather than relying on what the synthesized PIF XML happened to reference) according to Spoke-Hub architecture style. Sets app `root_agent` to the `is_root` group. Idempotent — safe to re-run. CXAS Version `0.0.5`. | ~10 sec | Updated `<target>_ir.json` stage history; CXAS app's `child_agents` set per group |

State flows through `<target>_ir.json` (a Pydantic `IRBundle` containing the `MigrationConfig`, source `DFCXAgentIR`, target `MigrationIR`, stage history, and version checkpoints). Each stage loads it from disk, mutates it, and writes it back. **No re-fetching or re-compiling between stages.**

## When to use this skill vs. the CLI directly

This skill (InquirerPy prompts + HTML pre-flight preview + Gemini model picker) is the right entry when you want to **interactively** drive a migration with rich pre-flight context. The same `MigrationService.run_stage*` methods this skill calls are also exposed via the canonical CLI for scripted / CI use:

```bash
# Same E2E plumbing, non-interactive (standard optimized profile by default):
cxas migrate dfcx --run --source-agent-id … --project-id … --target-name …

# Non-interactive Stage Checkpoint optimization runs:
cxas migrate dfcx --optimize --stage 1 --target-name my_app
cxas migrate dfcx --optimize --stage 2 --target-name my_app
cxas migrate dfcx --optimize --stage 3 --target-name my_app --architecture hub-and-spoke
cxas migrate dfcx --optimize --stage resume --target-name my_app  # interactive stage picker
```

The skill, the dashboard, and the E2E / Checkpoint commands all go through the **same** `MigrationService.run_stage_1/run_stage_2/run_stage_3` methods — pick whichever entry point matches your workflow.

## Prerequisites

```bash
# Ensure cxas_scrapi is installed editable (so this skill picks up local changes)
pip install -e .

# Auth
gcloud auth application-default login
gcloud auth list   # confirm the account has read on source + admin on target
```

InquirerPy is required for the interactive prompts (matches the agent-foundry skill):

```bash
pip install InquirerPy
```

## Driving the flow interactively (from Claude)

When invoked through Claude, lead the user through one question at a time. The scripts will prompt for missing inputs via InquirerPy, but you should pre-collect:

1. **GCP project ID** (target — where the new CXAS app will live).
2. **Location** — default `us`. **Do NOT default to `global` — it does not work for CXAS apps in most projects.**
3. **Source agent** — DFCX agent ID (full resource name) or path to a local `.zip` export.
4. **Target name** — display name for the new CXAS app.

Optional follow-ups: `--env` (PROD/AUTOPUSH), `--model` (Gemini), `--migration-version` (1.0/2.0).

## Quick Reference

```bash
# 1:1 migration (interactive — InquirerPy will prompt for project + location)
python .agents/skills/cxas-dfcx-migration/scripts/migrate.py

# Fully scripted
python .agents/skills/cxas-dfcx-migration/scripts/migrate.py \
  --source-agent-id "projects/<src_proj>/locations/us/agents/<uuid>" \
  --project-id <target_proj> --location us \
  --target-name my_cxas_app --yes

# Pre-flight HTML preview only (no migration)
python .agents/skills/cxas-dfcx-migration/scripts/migrate.py \
  --source-agent-id "<id>" --project-id <proj> --target-name preview_only \
  --preview-only --yes

# Stage 1 — variable dedup + Gemini consolidation
python .agents/skills/cxas-dfcx-migration/scripts/stage_1.py --target-name my_cxas_app

# Stage 1 — replay a saved grouping JSON
python .agents/skills/cxas-dfcx-migration/scripts/stage_1.py \
  --target-name my_cxas_app --grouping-json my_cxas_app_grouping.json --yes

# Stage 2 — instruction state machines + tool mocks + lint + report
python .agents/skills/cxas-dfcx-migration/scripts/stage_2.py --target-name my_cxas_app

# Stage 3 — rewire consolidated agent parent-child topology (idempotent)
python .agents/skills/cxas-dfcx-migration/scripts/stage_3.py --target-name my_cxas_app --architecture hub-and-spoke
```

## What lives in the skill vs. in `cxas_scrapi`

The skill is a thin orchestrator. **Every migration / optimization step lives in `src/cxas_scrapi/migration/` and is reachable via `MigrationService.run_stage_*` methods**, so the same logic powers all three entry points: this skill, `cxas migrate dfcx` (interactive TUI), and non-interactive command modes (`--run` / `--optimize`).

| Operation | src/ entry point |
|---|---|
| Source agent fetch / zip parse | `migration/dfcx_exporter.py:ConversationalAgentsAPI` |
| 1:1 migration | `migration/service.py:MigrationService.run_migration` |
| **Stage 1 orchestrator** (variable dedup + consolidation + integrity + topology link + orphan cleanup + versions + bundle persist) | `migration/service.py:MigrationService.run_stage_1` |
| **Stage 2 orchestrator** (state machines + tool mocks + unit-test regen + lint + audit report + bundle persist) | `migration/service.py:MigrationService.run_stage_2` |
| **Stage 3 orchestrator** (parent-child topology wiring) | `migration/service.py:MigrationService.run_stage_3` |
| Bundle persist convenience | `migration/service.py:MigrationService.persist_bundle` |
| Variable dedup primitive (Stage 1) | `migration/optimizer.py:CXASOptimizer.optimize_stage1` |
| Instruction restructuring + tool mocks primitive (Stage 2) | `migration/optimizer.py:CXASOptimizer.optimize_stage2` |
| Gemini N→M grouping + per-group PIF XML synthesis | `migration/structural_consolidator.py:StructuralConsolidator` |
| Pre-deploy integrity checks | `migration/integrity_checks.py:check_consolidation_integrity` |
| Parent-child topology + orphan cleanup | `migration/topology_wirer.py` |
| Update-pass redeploys | `migration/service.py:MigrationService._deploy_base_resources(is_update_pass=True)` + `_deploy_pending_agents(is_update_pass=True)` |
| Topology link | `migration/cxas_topology_linker.py` |
| Version checkpoints | `core/versions.py:Versions.create_version` |
| Topology SVG | `migration/graph_visualizer.py:HighLevelGraphVisualizer` |
| Per-resource Rich trees | `migration/playbook_visualizer.py` + `migration/flow_visualizer.py` |
| Deterministic unit tests | `migration/eval_generator.py:DeterministicEvalGenerator` |
| Migration report | `migration/dfcx_migration_reporter.py` |
| Optimization audit report | `migration/optimization_reporter.py:OptimizationReporter` |
| Grouping review TUI (accept / re-propose / merge / split / rename) | `cli/grouping_review.py:interactive_review` |
| HTML pre-flight preview | `migration/html_preview.py` |
| Post-deploy linter | `migration/post_deploy_lint.py` |
| IR bundle persistence | `migration/data_models.py:IRBundle` |

Skill-local helpers (UX glue only — InquirerPy prompts + thin
delegations to `MigrationCLI`):

- `_prompts.py` — InquirerPy prompt library (matches agent-foundry).
- `_shared.py` — InquirerPy variants of project/location prompts, source loader, and `MigrationConfig` assembly; plus pure delegations to `MigrationCLI` for `check_auth`, `run_dependency_analysis`, `select_resources`, `show_visualizations`.

The stage scripts now import the promoted modules directly:
``from cxas_scrapi.migration.data_models import IRBundle``
(plus ``html_preview`` in ``migrate.py`` and ``phase_tracker``) and call sites use the
canonical model names (``IRBundle``, ``phase_tracker.PhaseTracker``,
``html_preview.generate_html_report``) — no re-export shim layer.

The skill's stage scripts (`migrate.py` / `stage_1.py` / `stage_2.py` / `stage_3.py`) are now ~80-200 line shells: parse args → restore service from bundle → call the matching `MigrationService.run_stage_*` → print summary. There is no orchestration logic left in the skill — only InquirerPy prompts and the HTML preview that's specific to the skill's pre-flight UX.HTML preview that's specific to the skill's pre-flight UX.

## IR bundle (`<target>_ir.json`)

The unit of state shared across the three scripts. Pydantic `IRBundle` model:

```jsonc
{
  "schema_version": "1",
  "created_at": "2026-05-14T15:30:00",
  "config": { /* MigrationConfig */ },
  "source_agent_data": { /* DFCXAgentIR — needed for tool-mock context */ },
  "ir": { /* MigrationIR — mutated by each stage */ },
  "stage_history": [
    {"phase": "migrate", "status": "ok", ...},
    {"phase": "stage1",  "status": "ok", ...}
  ],
  "app_url": "https://ces.cloud.google.com/...",
  "version_checkpoints": [["0.0.1", "Stage 1: ..."]],
  "grouping": { /* present if Stage 1 ran consolidation */ }
}
```

Killing a stage script mid-run leaves the bundle untouched (only persisted on success). Re-running picks up where the last successful stage left off.

## Pre-flight HTML preview

`migrate.py` generates `<target>_tree_preview.html` in ~5 seconds after source loading. Open it in any browser to see:

- Source overview (resource counts, estimated migration time).
- Topology graph (graphviz SVG when `dot` is on PATH; Mermaid fallback otherwise).
- Per-playbook and per-flow Rich trees.

`migrate.py --preview-only` exits after the preview without running the migration.

## Troubleshooting

- **`create_app` returns `404 / 501 / MethodNotImplemented`** — your `--location` is wrong. CXAS apps in most projects live in `us`, not `global`. Pass `--location us`.
- **`AlreadyExists: App with same display name`** — pick a different `--target-name`. Old runs leave deployed apps behind even on partial failure.
- **Stage 1 / Stage 2 fail with `No IR bundle found`** — run `migrate.py` first to produce `<target>_ir.json`, or pass `--ir-bundle <path>` explicitly.
- **Synthesis (Stage 1 consolidation) hangs on Gemini** — fixed with per-group `asyncio.wait_for(timeout=600s)` in `structural_consolidator.synthesize_instructions`. Override via `SYNTHESIS_TIMEOUT_S` env var. Hung groups fall back to the concatenated instruction.
- **`gemini-2.5-flash-001 not found`** during AI augment — the Gemini call uses `locations/global` for the model; sometimes that endpoint is project-restricted. The migration continues with empty AI descriptions. Pick a different `--model` if you need them.

## Detailed reference

See [references/migration-options.md](references/migration-options.md) for full parameter / flag descriptions and the IR bundle schema.