Guide · Agent skill

How to use the Retiring and Inheriting a Seat agent skill

Updated 2026-11-02 · written with AI from the project's own documentation and checked against our catalogue data

This is an agent skill for OpenRig, a platform for running long-lived multi-agent setups. It guides an AI agent through a planned handover when its own working session nears its context limit: instead of letting compaction degrade it, the agent retires deliberately and hands its seat to a fresh successor. After this guide you will know what the skill makes the agent do, how to install it, how to trigger it with plain requests, and what it needs before you rely on it.

Retiring and Inheriting a Seatmvschwarz/openrig · Apache-2.0 Download the skill

Quick start

1Install OpenRig and the sibling skills. Make sure the rig command-line tool is available, and that seat-continuity-and-handover and session-compaction-and-restore are installed alongside this skill.
2Install this skill. Unzip the downloaded ZIP into ~/.claude/skills/, into a project's .claude/skills/ folder, or upload it under Settings → Capabilities → Skills in the Claude apps.
3Trigger it at a planned threshold. When the agent nears its context limit and a transition is planned, ask for a handover packet in plain language.
4Compose the packet and recap. Run rig context compose for the packet and rig context recap-write with your rig, seat and draft file.
5Prime the successor. Use rig walk to deliver the packet to the fresh successor, so it reads the handover as inheritance, not identity.
6Cut over and record the tenure. After the owner gives the go-ahead, keep the tmux seat, write the lineage ledger row and tombstone, and confirm the successor's self-check.

What the skill makes the agent do

The skill covers a planned seat transition. A seat is a role in a multi-agent setup, and its address stays stable while the agent occupying it changes over time. As a sitting agent approaches roughly 85% of its context window, the skill has it write a handover packet, record a row in an append-only lineage ledger, and help prime a fresh successor to take over the same seat.

The skill also covers a staged apprentice period before cutover, preserving the physical tmux seat (the canonical session, window and pane) at cutover, the framing that a successor inherits a seat rather than the predecessor's identity, and consulting a retired predecessor.

Handover packet: current work, next owner, durable pointers, constraints and authority boundaries, plus lineage notes from earlier occupants.
Lineage ledger: one row per tenure with generation, harness session id, start and retire times, packet pointer and a one-line tombstone.
Apprentice period: the successor observes and works in the real domain while the incumbent keeps authority until the named owner gives the go-ahead.
Consulting a predecessor: a bounded question to the retired agent, using the retained session.

Install the skill from the ZIP or the Claude apps

Download the skill ZIP from the directory page. The skill folder contains SKILL.md and a CHANGELOG.md. For Claude Code, unzip the folder into your user skills directory ~/.claude/skills/, or into a project's .claude/skills/ folder if you want the skill limited to one project.

In the Claude apps, you can instead upload the ZIP under Settings → Capabilities → Skills. Whichever route you use, the agent only gains the skill when it can see the folder, so check that the skill appears in your setup before relying on it.

Personal install: ~/.claude/skills/ (available across projects).
Project install: .claude/skills/ inside the project folder.
Claude apps: Settings → Capabilities → Skills, then upload the ZIP.

Prerequisites: what you need before it works

The skill assumes an existing OpenRig installation with the rig command-line tool, because its commands and seat concepts come from that platform. It also depends on two sibling skills, seat-continuity-and-handover and session-compaction-and-restore, which must be available alongside it. tmux is needed because the handover refers to the canonical tmux session and pane.

The skill is written for a seat that already has an occupant. If you are starting a fresh seat with no prior agent, the skill says to use rig launch or agent-starters instead.

OpenRig installed, with the rig command-line tool available.
The seat-continuity-and-handover and session-compaction-and-restore skills installed.
tmux, since the handover preserves the canonical tmux session, window and pane.
A seat that already has a predecessor. Fresh seats are not covered.

How to trigger the skill with natural requests

You do not need special syntax. The skill is meant to be picked up when an agent is near its context threshold and a planned transition is due, or when someone is priming a successor, recording a tenure, inheriting a seat, or asking about a previous occupant. Phrase the request around that situation so the agent recognises it.

If you want to consult the agent that held the seat before, say so directly. The skill also expects the agent to check whether an unplanned crash or compaction happened, because those cases belong to other skills.

'I am near my context limit and the seat handover is planned. Prepare a handover packet for my successor.'
'Record a lineage ledger row for this tenure and write my tombstone.'
'I am inheriting this seat. Explain the do-not-over-inherit framing and what I should carry forward.'
'Ask the retired predecessor of this seat about the migration it started.'

Run the handover sequence step by step

The handover follows a fixed order. The trigger comes first, either the configured continuity threshold or a deliberate role change; a context estimate alone does not authorize a cutover. The agent then composes the packet with rig context compose, writes the seat's recap, and primes the successor with rig walk or a launch with the packet. Before cutover, an apprentice period can be used for questions and domain work, while the incumbent keeps authority until the owner gives the go-ahead.

At cutover, the physical seat is preserved, and the agent writes the lineage row and tombstone and confirms the successor's self-check before handing over authority.

Trigger: the configured threshold or a planned role transition, under the named transition owner.
Author the packet and enumerate standing duties, including who holds each after cutover.
Write the recap: run rig context recap-write with the rig, seat and draft file.
Prime the successor: rig walk the packet in, or launch with it. The packet's first line points to orienting-to-an-inherited-seat.
Assess before cutover, then preserve the tmux seat and write the ledger row and tombstone.

rig context compose
rig context recap-write --rig <rig> --seat <seat> --file <draft>
rig walk

Ask a retired predecessor with rig ask --wake

The wake surface gives bounded consultation of a retained session. The skill says the packet should include the exact resume handle and the questions only the predecessor can answer. The predecessor's answers are testimony, not authoritative state, and the skill asks the agent to record where the old history lives and its known limits.

If the consultation fails, the skill asks you to distinguish missing history, runtime failure, timeout and context exhaustion. Check rig ask --help for runtime selection and timeout options, since the skill does not list them.

Use rig ask --wake for a bounded question to a retained session.
Treat answers as testimony to check, not as proof of current state.
Use a supported runtime resume path only when its prerequisites and the requested authority allow it.

rig ask --wake
rig ask --help

Safety notes: scripts and network

The skill itself runs no scripts and uses no network. It is a set of written instructions that the agent follows, and the commands it refers to, such as rig context compose and rig walk, belong to the OpenRig tool. Review those commands against your own OpenRig setup before running a handover on a live seat.

Because a handover changes which agent holds authority, read the authority boundaries in the packet before cutover. The skill says the incumbent keeps authority until the named owner gives the go-ahead, and the successor should not act as the seat before then.

No scripts are bundled with the skill.
The skill does not use the network itself.
The handover moves authority, so confirm the owner and the cutover word before cutover.

What people use it for

Replacing a long-running coordinator agent before its context fills, so the successor starts clean with the packet.
Keeping an auditable record of which agent held a seat and during which period, using one ledger row per tenure.
Transferring recurring duties, such as a regular report, explicitly to the successor at cutover.
Asking a retired predecessor a specific question about work it started, using rig ask --wake.

Frequently asked questions

Is the retiring and inheriting a seat skill free to use?

The source does not state pricing or licence terms beyond the Apache-2.0 licence listed for the repository, which applies to the code in the OpenRig project. Check the licence file in the repository before redistributing or modifying it.

What do I need installed before this skill works?

You need an OpenRig installation with the rig command-line tool, the seat-continuity-and-handover and session-compaction-and-restore skills, and tmux. The skill is for seats that already have a predecessor.

Where do I put the skill folder?

Unzip it into ~/.claude/skills/ for personal use, or into a project's .claude/skills/ folder. In the Claude apps, upload the ZIP under Settings → Capabilities → Skills.

Does this skill handle crashes or automatic compaction?

No. The skill says unplanned compaction or crash recovery belongs to session-compaction-and-restore and claude-compaction-restore. This skill is for planned transitions only.

Can I use this skill for a brand-new agent seat?

No. The skill says a fresh seat with no occupant should use rig launch or agent-starters instead.

Does the skill run scripts or use the network?

According to the skill listing, it runs no scripts and uses no network. The rig commands it references come from OpenRig, so review those separately.

New agent skills, repos and AI tools — every week

The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.

Our verdict

This skill suits developers and teams running long-lived multi-agent setups on OpenRig who want planned, auditable handovers and a successor that starts from explicit authority boundaries. If you do not use OpenRig, have no tmux setup, or need crash or automatic compaction recovery, pick the sibling skills or a different tool instead.

Download the skill