Intrinsic Core API Overview

What it does

When your agent uses it

• Tasks that involve gRPC and Protobuf
• Tasks that involve Cloud networking
• Tasks that involve Container orchestration

Example prompts

• “/intrinsic-core-api-overview”

Requirements

• Python 3

Workflow steps

5 steps, taken from the first numbered list in SKILL.md.

1. 1Envoy ingress routing headers: Attach x-resource-instance-name: metadata header when calling deployed asset instances (e.g. icon…
2. 2Context handle access across execution phases: Read leased equipment connections via context.resource_handles[""] strictly within…
3. 3Spatial pose construction parameter ordering: Supply explicit keyword arguments when instantiating Pose3(rotation=..., translation=...)…
4. 4Multi-world transformation scoping and cleanup: Specify node_to_update when invoking update_transform across non-neighboring nodes, and…
5. 5CLI addressing and subcommand selection: Supply --address=localhost:17080 for commands (run inctl service state list…

What it can do on your machine

Read from SKILL.md and the folder at commit 0221644. It shows what the files ask for, not the result of running them.

• Tool permissions

  Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

  From allowed-tools in the SKILL.md frontmatter.

• Runs code

  Shell commands in SKILL.md call:

  • bazel
  • kubectl

  From the folder's file list and the shell code blocks in SKILL.md.

• Network

  No URLs in SKILL.md. Its commands use kubectl, which can reach the network depending on how they are called.

  From URLs in SKILL.md, links to its own repository left out.

• Credentials

  Names no API keys, tokens, secrets or passwords.

  From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Intrinsic Core API Overview loads about 3.4k tokens when it runs, and up to ~83k if it reads all its reference files. Until then it costs about 195 tokens; SKILL.md has 792 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it

~195

When it runs · the whole SKILL.md, loaded when a task matches

~3.4k

With references · SKILL.md plus every file in references/, read only if the agent opens them

~83k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from intrinsic-ai/intrinsic-core at commit 0221644, republished under its Apache-2.0 licence (© intrinsic-ai). 792 words, ~3,400 tokens.

name

intrinsic-core-api-overview

description

Intrinsic Core gRPC service selection, Envoy x-resource-instance-name routing, and progressive disclosure hub across ObjectWorldService, MotionPlannerService, ICON, cameras, KVStore, and platform APIs. Triggers: select Intrinsic Core API, gRPC routing, Envoy ingress, x-resource-instance-name, asset instance vs platform service, coordinate frame mutation, multi-world lifecycle, trajectory planning, real-time control, camera capture, KVStore. Subsystems: INGRESS, ENVOY, WORLD, MOTION_PLANNING, ICON, PERCEPTION, EXECUTIVE, KVSTORE. Dependencies: localhost:17080, grpc, intrinsic.util.grpc, ObjectWorldClient, MotionPlannerClient, icon_api. Anti-keywords: inctl service create, inctl skill create, inctl solution list, kubectl scale, legacy WORKSPACE.

import os
from typing import Any
import grpc
from intrinsic.util.grpc import connection, interceptor
from intrinsic.world.proto import object_world_service_pb2_grpc
from intrinsic.world.python import object_world_client

_CHANNEL_OPTIONS = [("grpc.max_receive_message_length", -1)]


def resolve_workcell_ingress_address(default_address: str = "localhost:17080") -> str:
  """Resolves in-cluster Envoy ingress DNS when executing inside a Kubernetes pod."""
  if "KUBERNETES_SERVICE_HOST" in os.environ:
    return "istio-ingressgateway.app-ingress.svc.cluster.local:80"
  return default_address


def connect_to_asset_instance(
    instance_name: str, address: str | None = None
) -> grpc.Channel:
  """Creates a gRPC channel routed to a specific named Intrinsic asset instance."""
  params = connection.ConnectionParams(
      address=address or resolve_workcell_ingress_address(),
      instance_name=instance_name,
      header="x-resource-instance-name",
  )
  base = grpc.insecure_channel(params.address, options=_CHANNEL_OPTIONS)
  return grpc.intercept_channel(
      base, interceptor.HeaderAdderInterceptor(params.headers)
  )


def create_channel_from_resource_handle(handle: Any) -> grpc.Channel:
  """Extracts connection_info.grpc from a leased ResourceHandle message."""
  info = handle.connection_info.grpc
  return connect_to_asset_instance(
      instance_name=info.server_instance,
      address=info.address or resolve_workcell_ingress_address(),
  )


def connect_to_platform_world_service(
    address: str | None = None, world_id: str = "world"
) -> object_world_client.ObjectWorldClient:
  """Connects directly to the singleton ObjectWorldService without instance headers."""
  channel = grpc.insecure_channel(
      address or resolve_workcell_ingress_address(), options=_CHANNEL_OPTIONS
  )
  stub = object_world_service_pb2_grpc.ObjectWorldServiceStub(channel)
  return object_world_client.ObjectWorldClient(world_id=world_id, stub=stub)

[gRPC / CLI error during Intrinsic Core API workflow]
  │
  ├─► [Symptom: rpc error: code = Unimplemented desc = (empty description)]
  │     └─► Cause: Missing x-resource-instance-name header on asset instance or invalid header on platform service.
  │     └─► Action: Add x-resource-instance-name for asset instances, or omit header for core platform services.
  │
  ├─► [Symptom: rpc error: code = Unavailable desc = no healthy upstream]
  │     └─► Cause: Target container crashed (SIGSEGV), inference sidecar exited (128), or CFS CPU throttling.
  │     └─► Action: Check kubectl describe pod in app-intrinsic-app-chart; verify upstream service dependencies.
  │
  ├─► [Symptom: grpc: received message larger than max (X vs 4194304)]
  │     └─► Cause: Mesh or point cloud transfer exceeded default 4 MB gRPC receive message ceiling.
  │     └─► Action: Configure channel with [("grpc.max_receive_message_length", -1)] or 64 MB limit.
  │
  ├─► [Symptom: KeyError: 'resource_handles' or AttributeError during Skill.preview()]
  │     └─► Cause: Attempting to access live hardware resource handles during preview phase.
  │     └─► Action: Use context.get_object_for_equipment("<slot>") and record updates via context.record_world_update().
  │
  ├─► [Symptom: INVALID_ARGUMENT: node_to_update is required when updating non-neighboring nodes]
  │     └─► Cause: update_transform invoked across multi-hop transform path without disambiguating target node.
  │     └─► Action: Pass node_to_update=target_node explicitly to update_transform().
  │
  ├─► [Symptom: HTTP 404 Not Found on REST gateway /api/kvstore/stores/...]
  │     └─► Cause: Envoy gateway missing mandatory /api/http-gateway/ path prefix rewrite.
  │     └─► Action: Prepend /api/http-gateway to the request path (/api/http-gateway/api/kvstore/...).
  │
  └─► [Symptom: failed to upload image: failed to dial "/run/containerd/containerd.sock": connect: connection refused]
        └─► Cause: Backend artifacts-deployment pod in k3s cannot reach host containerd socket.
        └─► Action: Cap at <= 2 attempts; do not inspect local /run or proxy socket; verify build/test hermetically with bazel.

© intrinsic-ai, Apache-2.0. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 7 other files (references) in .agents/skills/intrinsic-core-api-overview of intrinsic-ai/intrinsic-core.

• SKILL.md
• references/assets-and-solutions.md
• references/geometry-and-math.md
• references/longrunning-operations.md
• references/motion-planning-and-icon.md
• references/perception-and-hardware.md
• references/platform-logging-and-status.md
• references/world-and-kinematics.md

Open the folder on GitHubat commit 0221644