handoff

handoff is a skill for Claude Code from spacegrowth/claude-relay. It costs 44 tokens per session (830 once invoked), scanned A, original, MIT.

A handoff skill that prepares a short summary for a new coding-agent session and starts that successor session. A handoff transfers current work, decisions, and pending tasks when a conversation becomes too large.

In plain words
What is it for?
Use it when asked to hand off work or when a long-running lead session needs a fresh successor. It writes a handoff file and invokes the project's relay command.
Why use it?
It preserves important context without carrying an unwieldy conversation into the next session. The summary records open questions, next steps, and assigned responsibility.

Skill for Claude Code

Written for Claude Code: ${CLAUDE_PLUGIN_ROOT} variable. Also seen: mentions Claude Code.

Runs only inside its plugin — its command needs a path that Claude Code sets for a plugin’s own hooks and for nothing else. Install the plugin, not this.

Part of the relay plugin — 18 skills, 1 agent, 4 hooks shipped together

Good fit Use it when asked to hand off work or when a long-running lead session needs a fresh successor. It writes a handoff file and invokes the project's relay command.

Compare 6 skills from other repositories ↓
Install

Getting it into your agent

This one installs as part of its plugin. Adding the marketplace and installing the plugin brings it with everything else the plugin ships.

Claude Code
/plugin marketplace add spacegrowth/claude-relay
Claude Code
/plugin install relay

Made for: Claude Code.

Or install relay, the plugin that ships this one along with the rest of its 18 skills, 1 agent, 4 hooks.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for handoff

README.md
[![agentmods](https://agentmods.dev/badge/skills/spacegrowth/claude-relay/handoff/github.svg)](https://agentmods.dev/skills/spacegrowth/claude-relay/handoff)
Your own site
<a href="https://agentmods.dev/skills/spacegrowth/claude-relay/handoff"><img src="https://agentmods.dev/badge/skills/spacegrowth/claude-relay/handoff/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for handoff

Your own site · 80×15
<a href="https://agentmods.dev/skills/spacegrowth/claude-relay/handoff"><img src="https://agentmods.dev/badge/skills/spacegrowth/claude-relay/handoff.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 44 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 830 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.00044 $0.00830
Opus 5 $0.00022 $0.00415
Sonnet 5 $0.00009 $0.00166
Haiku 4.5 $0.00004 $0.00083

Measured yesterday against content hash 22e9bd41fe4a, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-08, from the pricing page.

Security

Grade A, and why

handoff scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

skills/handoff/SKILL.md · 55 lines

What it actually says

Use this when the lead session has gotten heavy (long-lived, large transcript — see the auto-wake handoff nudge) and the healthy move is a fresh context, not more of this one.

First, write the handoff file — a short packet-file-style summary under ~/.relay-tasks/ (gate-exempt), covering: what's in flight, what's reviewed/committed, open questions, next steps. Write it like a memo to your successor, not a transcript dump. Carry forward every [discipline] marker your own predecessor's doc had (relay warns if one goes missing — see the linter note below), and on every queue item name an EXECUTOR to do it (or write "delegate this") — never phrase box/deploy/ops work as this lead's own to pick up (§10's [ops-not-lead-work] incident: a dropped marker plus ops phrased as the lead's task is exactly how a successor drifted into implementation).

Then run:

${CLAUDE_PLUGIN_ROOT}/bin/relay handoff <path-to-handoff.md>

Call relay as ${CLAUDE_PLUGIN_ROOT}/bin/relay (Claude Code substitutes the plugin's absolute path when this skill loads) — not bare relay, which often isn't on the Bash tool's non-interactive PATH. Unlike other relay commands, handoff reads the caller's session id from $CLAUDE_CODE_SESSION_ID automatically — no need to pass it yourself.

This opens a NEW lead tab, pre-armed (its routing gate and auto-wake are already active from turn one), seeded with a short pointer at a relay-prepared copy of the handoff file — your source md is never modified; relay appends a SUCCESSOR AFTERCARE section to its own copy instead, so the successor's launch prompt stays a one-line pointer (a prior version inlined the aftercare directly into the prompt and it truncated a real launch command — see build_handoff_copy). As its final act, this steps the CURRENT session down — the gate and auto-wake here turn off. Any executors this lead owned are re-parented to the successor here, and each one's already-seen report stamps travel with it, so a report the outgoing lead reviewed and committed doesn't re-wake the successor as new. If any of them are pinned (relay keep / spawn --keep), the handoff output names them — auto-close will never park a pinned executor, so the successor inherits them indefinitely until someone runs relay keep <sid> --off.

Aftercare, for the successor. Once it's settled in (read the handoff, run /relay:list), the successor runs /relay:mode — idempotent if the pre-arm pin held, and it arms for real if Claude Code minted a different session id for the tab. It then asks the user whether to close the predecessor's now-unarmed tab, and on a yes runs relay close-predecessor (never unasked) — the predecessor's tab identity travels in the successor's marker for exactly this.

Optional flags: --project NAME and --model NAME override the successor's project/model (default: inherited from this lead's own marker).

Linter note. relay handoff compares the doc THIS lead itself inherited against the doc it's handing off and warns if any [discipline] marker went missing (dropped_discipline_markers) — silent when nothing was dropped, or when this is the first-ever lead with no inherited doc to compare against.

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. yesterday Changed · +3 lines 22e9bd41fe4a
  2. 8d ago First seen · 52 lines · 44 tokens per session scan A 0e242e1f2ab7

Subscribe to this mod's changes

handoff is a skill published in the GitHub repository spacegrowth/claude-relay (1 stars, last pushed yesterday), licensed MIT. It adds 44 tokens to every session and 830 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.