v0.6.0 to v0.8.0

66 added, 80 removed. Audit A to A.

---

name: i4h-workflow

- version: "0.6.0"

- description: Overview of `workflows/agentic/` (IsaacLab-Arena + GR00T/openpi). Use when the user asks what i4h workflow is, what's supported, or where to start.

+ description: Orient users to the i4h workflow runtime and route them to the correct stage skill. Use for architecture, support, or where-to-start questions; do not execute a known stage.

license: Apache-2.0

metadata:

author: "Isaac for Healthcare Team <isaac-for-healthcare-support@nvidia.com>"

+ version: "0.8.0"

tags:

- isaac-for-healthcare

- i4h

- - agentic-workflow

- robotics

- - overview

+ - onboarding

---

- # i4h Agentic Workflow

+ # i4h Workflows

## Purpose

- Orient on the agentic workflow before touching a specific stage: which envs/robots/policies are supported, how the `workflows/agentic/` subprojects fit together, and which per-stage skill to invoke next. This skill routes — it runs no pipeline stage itself.

+ Orient the user from live repository facts, then hand execution to the narrowest stage skill.

- ## Base Code

+ ## Instructions

- These steps drive the i4h-workflows base code (the `workflows/agentic/` tree). To reuse an existing checkout, set `I4H_WORKFLOWS` to its path (no clone happens). Otherwise this resolves the current repo, or clones to `~/i4h-workflows` — pick that default without prompting. Run every command below from the resolved root:

+ 1. Run the base-checkout resolver.

+ 2. Read live support and `DESIGN.md`.

+ 3. Use only current architecture facts in the answer.

+ 4. Use the narrowest stage skill for execution.

+ ## Resolve the checkout

```bash

- # Resolve the i4h-workflows base code (provides workflows/agentic/).

+ export I4H_WORKFLOWS_REPO_URL="${I4H_WORKFLOWS_REPO_URL:-https://github.com/isaac-for-healthcare/i4h-workflows}"

+ I4H_REPO_DIR_NAME="${I4H_WORKFLOWS_REPO_URL%/}"

+ I4H_REPO_DIR_NAME="${I4H_REPO_DIR_NAME##*/}"

+ I4H_REPO_DIR_NAME="${I4H_REPO_DIR_NAME##*:}"

+ I4H_REPO_DIR_NAME="${I4H_REPO_DIR_NAME%.git}"

+ [ -n "$I4H_REPO_DIR_NAME" ] || { echo "Cannot derive a checkout name from I4H_WORKFLOWS_REPO_URL" >&2; exit 2; }

ROOT="${I4H_WORKFLOWS:-$(git rev-parse --show-toplevel 2>/dev/null)}"

- if [ ! -d "$ROOT/workflows/agentic" ]; then

- ROOT="${I4H_WORKFLOWS:-$HOME/i4h-workflows}"

- [ -d "$ROOT/workflows/agentic" ] || git clone https://github.com/isaac-for-healthcare/i4h-workflows "$ROOT"

+ if [ ! -d "$ROOT/workflows/i4h_workflows" ]; then

+ ROOT="${I4H_WORKFLOWS:-$HOME/$I4H_REPO_DIR_NAME}"

+ [ -d "$ROOT/workflows/i4h_workflows" ] || git clone "$I4H_WORKFLOWS_REPO_URL" "$ROOT"

- export I4H_WORKFLOWS="$ROOT"; cd "$ROOT"

+ export I4H_WORKFLOWS="$ROOT"

+ cd "$ROOT"

```

- ## Basics

- - Env YAMLs at `workflows/agentic/config/environments/<env>.yaml` are the source of truth.

- - Each pipeline stage has its own skill. Compose them or use [[i4h-workflow-e2e]] for full runs.

- - For env authoring or scene-edit source changes, load `skills/i4h-workflow/references/repo-map.md` before generating code.

+ Treat this resolver as part of the skill contract: a hosted copy may run outside the base repository, so never assume the current checkout contains `workflows/i4h_workflows`. `I4H_WORKFLOWS_REPO_URL` selects the clone source. When `I4H_WORKFLOWS` is unset, derive the fallback directory from that URL; set `I4H_WORKFLOWS` only to reuse or choose a specific destination. Never replace an existing checkout.

- ## Supported Envs

+ ## Inspect before answering

- Fallback when `--list-envs` fails (setup incomplete):

+ Read `./DESIGN.md` for architecture and `skills/i4h-workflow/references/repo-map.md` for ownership. Discover current support instead of copying a static table:

- | Env | Robot | Policy |

- |---|---|---|

- | `scissor_pick_and_place` | SO-ARM 101 | GR00T N1.5 (N1.7 alternative) |

- | `locomanip_tray_pick_and_place` | Unitree G1 | GR00T N1.6 (shared `policy.locomanip.*`) |

- | `locomanip_push_cart` | Unitree G1 | GR00T N1.6 (shared `policy.locomanip.*`) |

- | `assemble_trocar` | Unitree G1 + Dex hands | GR00T N1.5 (inference-only) |

- | `ultrasound_liver_scan` | Franka-style arm | openpi PI0 |

- | `surgical_reach_psm` | dVRK PSM | GR00T N1.5 or scripted state machine |

- | `surgical_reach_dual_psm` | dVRK dual PSM | GR00T N1.5 or scripted state machine |

- | `surgical_reach_star` | STAR | GR00T N1.5 or scripted state machine |

- | `surgical_lift_block` | dVRK PSM | GR00T N1.5 or scripted state machine |

- | `surgical_lift_needle` | dVRK PSM | GR00T N1.5 or scripted state machine |

- | `surgical_lift_needle_organs` | dVRK PSM | GR00T N1.5 or scripted state machine |

+ ```bash

+ ./run.sh list

+ ```

- ## Run

+ If discovery fails because setup is incomplete, report that limitation and route to `i4h-workflow-setup`.

- Run the step below before answering. Each step is a separate bash call; variables persist in the local agent's tmux session.

+ ## Explain the design

- ### Step 1 — list supported envs

+ Keep the summary precise:

- ```bash

- REPO_ROOT="${I4H_WORKFLOWS:-$(git rev-parse --show-toplevel 2>/dev/null)}"; [ -d "$REPO_ROOT/workflows/agentic" ] || REPO_ROOT="$HOME/i4h-workflows"

- "${REPO_ROOT}/workflows/agentic/policy/run.sh" --list-envs

- ```

+ - A Scene owns the simulated world, assets, embodiment, cameras, randomization, adapters, and reset hooks.

+ - A Task owns one reusable capability. It reads `ctx.scene`, writes `ctx.act`, and never advances the simulator.

+ - A Workflow selects one Scene, exposes run-mode-specific `TaskGraph` builders, and owns goal semantics. A run mode answers how that workflow should run; code and CLI use the shorter term `mode`.

+ - The Engine schedules graph nodes; the shared `SimulationRunner` alone resets, steps, renders, records, retries whole episodes, and prints run summaries.

+ - Online RL is a separate training lifecycle: its trainer owns vectorized stepping and returns a checkpoint to the normal policy Task and `SimulationRunner` validation path.

+ - Simulator-compatible exported RSL-RL actors may run as in-process Tasks; incompatible foundation-model policy stacks remain remote.

+ - Remote policy stacks run out of process and communicate over Zenoh; offline dataset tools remain independent of the simulator.

+ - Python owns behavior. Manifests carry facts across dependency boundaries.

- If the command fails, use the **Supported Envs** fallback table above.

+ Do not describe retired environment YAMLs, per-mode runners, or separate policy/Arena launchers.

- ## Subprojects

+ ## Route the next action

- | Directory | Purpose |

+ | Goal | Skill |

|---|---|

- | `workflows/agentic/arena/` | IsaacLab-Arena envs, scenes, tasks, teleop, record, replay |

- | `workflows/agentic/policy/` | Policy daemons and train dispatchers |

- | `workflows/agentic/dataset/` | HDF5 → LeRobot conversion and visualization |

- | `workflows/agentic/mimic/` | HDF5 trajectory expansion |

- | `workflows/agentic/annotator/` | VLM success labels and filtering |

- | `workflows/agentic/cosmos/` | Optional Cosmos Transfer video augmentation |

- | `workflows/agentic/common/` | Shared config, messaging, robot constants |

+ | Install, sync, or repair dependencies | `i4h-workflow-setup` |

+ | Create a new workflow/environment | `i4h-workflow-create` |

+ | Edit an existing scene, camera, task, or success rule | `i4h-workflow-scene-edit` |

+ | Record demonstrations | `i4h-workflow-dataset-teleop` |

+ | Replay HDF5 | `i4h-workflow-dataset-replay` |

+ | Augment HDF5 | `i4h-workflow-dataset-mimic` |

+ | Grade/filter HDF5 with a VLM | `i4h-workflow-dataset-annotate` |

+ | Convert HDF5 to LeRobot | `i4h-workflow-dataset-convert` |

+ | Inspect LeRobot in a browser | `i4h-lerobot-viz` |

+ | Fine-tune a manifest-backed policy task | `i4h-workflow-finetune` |

+ | RL post-train a supported policy in simulation | `i4h-workflow-train-rl` |

+ | Run policy or rule-based rollouts | `i4h-workflow-validate` |

+ | Run the maintained complete pipeline | `i4h-workflow-e2e` |

- ## Skill Index

+ For `Stop all`, do not load a stage skill. Run `./stop.sh all` from the repository root and report the stopped process count.

- | Skill | Purpose |

- |---|---|

- | `i4h-workflow-setup` | Install / sync / check third-party deps |

- | `i4h-workflow-create` | Add a new env |

- | `i4h-workflow-scene-edit` | Edit an existing scene / task / camera |

- | `i4h-workflow-dataset-teleop` | Record human demos |

- | `i4h-workflow-dataset-replay` | Replay HDF5 episodes |

- | `i4h-workflow-dataset-mimic` | Expand HDF5 demos with noise |

- | `i4h-workflow-dataset-annotate` | VLM label / filter episodes |

- | `i4h-workflow-dataset-convert` | Convert HDF5 to LeRobot |

- | `i4h-workflow-finetune` | Train supported envs |

- | `i4h-workflow-validate` | Roll out / evaluate policy checkpoints |

- | `i4h-workflow-e2e` | Run the full pipeline |

- | `i4h-lerobot-viz` | Open the LeRobot HTML viewer |

+ ## Troubleshooting

+ If discovery fails, verify the resolved checkout and run setup. If a mode is absent, report it as unsupported.

## Prerequisites

- - For any hands-on stage, set up the workflow first — see [[i4h-workflow-setup]] (component venvs and third-party checkouts present).

- - Env YAMLs at `workflows/agentic/config/environments/<env>.yaml` are the source of truth.

+ Require a readable base checkout or network access to clone it.

## Limitations

- - Overview/routing only — each pipeline stage has its own skill; this one performs no recording, training, or rollout.

- ## Troubleshooting

- - **Error:** env not found / unsupported — Cause: typo, unregistered env, or setup incomplete. Fix: run `workflows/agentic/policy/run.sh --list-envs`; if that fails, run [[i4h-workflow-setup]] first; otherwise check the **Supported Envs** table.

- ## Final Response

- The response is incomplete unless it includes all three parts below, each separated by a blank line. Do not reply with prose only or run sections together without spacing.

+ This router does not install, author, simulate, process data, train, or evaluate.

- 1. **Orientation** — one short paragraph: env YAMLs, subprojects, sim-to-policy pipeline.

+ ## Examples

- 2. **Supported envs** — heading plus markdown table from Step 1 (`--list-envs` output: env, stack, description) or the **Supported Envs** fallback table.

+ - `What does the i4h workflow include, and where should I start?` → inspect live support, summarize `DESIGN.md`, and recommend one stage skill.

- 3. **Available skills** — heading plus the full **Skill Index** table copied into the response.

+ ## Completion gate

- Then (after another blank line) recommend [[i4h-workflow-setup]] if setup is not done, and name one next stage skill matched to the user's goal.

+ Answer with the live workflow/mode list, a short architecture summary, and one concrete next skill. If the requested workflow or mode is absent from `run.sh list`, say it is unsupported instead of inventing a command.