Planning Held Object Motion logo

Planning Held Object Motion

Organization
graph-robots
planning-held-object-motion

Plans the carry and engagement phases for an already-held rigid object: a clearance-first lift, the smallest feasible symmetry-equivalent reorientation, an orientation-locked transit above the fixture, or a direct plan to the approach pose; then a typed engagement or a straight linear insertion, optionally re-observing the held tip from the wrist first. Use when a held object must be brought to a fixture for insertion, hanging, packing, racking, or constrained sorting and an execution skill will run the resulting plans.

Overview

Publishergraph-robots
Repositoryopen-robot-skills
Skill nameplanning-held-object-motion
Stars
58
Forks
8
Bundled files
3
LicenseApache-2.0
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 3 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by graph-robots on GitHub. Read the source before you install it.

Installation

Install the Planning Held Object Motion AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/graph-robots/open-robot-skills/tree/main/skills/planning-held-object-motion
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/graph-robots/open-robot-skills.git /tmp/open-robot-skills
mkdir -p .claude/skills
cp -r /tmp/open-robot-skills/skills/planning-held-object-motion .claude/skills/planning-held-object-motion
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Planning Held Object Motion in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Planning Held Object Motion on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Planning Held Object Motion is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

planning-held-object-motion

Turn a feature mate into safe phases for a held object:

text
lift away from support
  -> rotate minimally in clear space
  -> translate above the fixture with orientation fixed
  -> approach and engage linearly

The skill plans; it does not move the robot or release the object. Follow it with executing-held-object-motion to execute the carry plan and executing-feature-mating to execute the engagement plan.

Choosing the carry strategy

plan_clearance_motion takes strategy:

  • "clearance_first" (default) derives a lift from the farthest attached sphere, selects the smallest reachable symmetry-equivalent rotation among eight roll candidates (each checked with motion.plan_joint), rotates while clear, and translates the aligned feature to a staging location on the free side of the fixture, approach_clearance_m along its axis. The staging location is derived from the fixture geometry and attached-object extent. stage_outward_m (default 0) shifts that location within the support plane away from the fixture toward the current hand — use it when a recessed opening is observed mostly at its far rim and the object should stage over the interior instead.
  • "direct" escapes escape_m (default 6 cm) along the support normal as a contact_transition waypoint, then hands approach_pose to the executor as one planned_joint waypoint with allow_start_contact and three attempts. It calls no planner itself. Use it when the object is already clear of its support, the orientation change is small, and a single attached-object plan to the approach pose is appropriate; it requires approach_pose.

If uncertain, choose the clearance-first strategy for a large orientation change or a long held object. Do not add arbitrary midpoint waypoints and do not remove collision geometry to make direct planning succeed.

Engagement

plan_feature_engagement converts the typed relation and the approach, engaged, and mate poses into execution semantics. loop_over_shaft crosses the shaft tip with a tracked cartesian_cross leg — stopping on first contact can leave a loop balanced against the tip without enclosing the shaft — and then seats with the bounded contact_seat servo; feature_to_fixture uses the contact servo for both legs; aperture insertion uses planned_linear waypoints that allow goal contact. The task graph therefore does not decide ad hoc whether a contact servo is needed.

plan_linear_engagement plans a straight insertion from the fixture feature alone: a planned_joint pre-contact pose precontact_clearance_m before the opening and a planned_linear engaged pose engagement_depth_m past it, both holding the current orientation. Give it observation (the current Observation), object_description and direction_marker_description to re-observe the held tip from the wrist first: when both the object and its direction marker are visible, the endpoint of the object's principal axis toward the marker replaces the carried feature position.

Recommended skill sequence

  1. Use registering-held-objects to obtain held_feature_in_tcp and attached_object.
  2. Use computing-feature-mating-poses when explicit approach/engaged/mate poses are needed.
  3. Select the carry strategy using the criteria above.
  4. Execute the returned reorientation_plan with executing-held-object-motion.
  5. Execute placement_plan with executing-feature-mating.

Waypoint modes are semantic: contact_transition leaves the initial support, planned_joint is collision-aware free-space motion, planned_linear locks orientation for a straight local leg, cartesian_cross is a tracked short crossing, and contact_seat is the final intentional-contact servo.

Boundaries

  • Upstream perception supplies the held feature in TCP coordinates and the fixture pose/axis. Registration supplies attached-object collision geometry.
  • Inputs describe functional geometry and a relation rather than an object class or task name; the graph chooses strategy, stage_outward_m and the re-observation descriptions.
  • Free roll about the fixture axis is treated as symmetry. Candidates are checked for reachability and the smallest feasible TCP rotation is selected.
  • support_normal may be supplied when the escape direction is known; it defaults to world up for a horizontal support surface.
  • Relationship-specific clearances may be supplied by the graph when fixture depth is observable or specified. Conservative geometric defaults are used otherwise.
  • The final crossing is linear and may allow goal contact. All earlier motion remains collision-aware free-space motion.

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Planning Held Object Motion AI skill do?

Plans the carry and engagement phases for an already-held rigid object: a clearance-first lift, the smallest feasible symmetry-equivalent reorientation, an orientation-locked transit above the fixture, or a direct plan to the approach pose; then a typed engagement or a straight linear insertion, optionally re-observing the held tip from the wrist first. Use when a held object must be brought to a fixture for insertion, hanging, packing, racking, or constrained sorting and an execution skill will run the resulting plans.

Why use Planning Held Object Motion on TypingMind?

Because you install it once and use it with any model. Planning Held Object Motion is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Planning Held Object Motion in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/graph-robots/open-robot-skills/tree/main/skills/planning-held-object-motion. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Planning Held Object Motion?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Planning Held Object Motion?

As many as you like. As long as a model supports skills, you can use Planning Held Object Motion with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Planning Held Object Motion AI skill free?

Yes. It is published on GitHub by graph-robots under the Apache-2.0 license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇