handoff

handoff is a skill for Claude Code from brunob54/superpowers-orchestrator. It costs 46 tokens per session (1,392 once invoked), scanned A, original, MIT.

A tool that saves the current work state as a continuation prompt in a temporary documentation file. It records details such as the branch, recent commits, changed files, and the next task so another session can continue.

In plain words
What is it for?
Use it before switching sessions, handing work to another agent, or preserving the state of an unfinished coding task.
Why use it?
It reduces the information lost when a context window is cleared or work moves to another session. The saved facts let the next agent continue without guessing what has already been done.

Skill for Claude Code

Written for Claude Code: disable-model-invocation in frontmatter. Also seen: mentions CLAUDE.md; mentions subagents.

Part of the superpowers-orchestrator plugin — 32 skills, 2 agents, 5 hooks shipped together

Good fit Use it before switching sessions, handing work to another agent, or preserving the state of an unfinished coding task.

Compare 6 skills from other repositories ↓
Install with agentmods
npx agentmods add skills/brunob54/superpowers-orchestrator/handoff
Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

Any agent
npx skills add brunob54/superpowers-orchestrator --skill handoff
Clone the repo
git clone --depth 1 https://github.com/brunob54/superpowers-orchestrator

Made for: Claude Code.

Or install superpowers-orchestrator, the plugin that ships this one along with the rest of its 32 skills, 2 agents, 5 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/brunob54/superpowers-orchestrator/handoff/github.svg)](https://agentmods.dev/skills/brunob54/superpowers-orchestrator/handoff)
Your own site
<a href="https://agentmods.dev/skills/brunob54/superpowers-orchestrator/handoff"><img src="https://agentmods.dev/badge/skills/brunob54/superpowers-orchestrator/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/brunob54/superpowers-orchestrator/handoff"><img src="https://agentmods.dev/badge/skills/brunob54/superpowers-orchestrator/handoff.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 46 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 1,392 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 $0.00046 $0.01392
Opus 5.5 $0.00018 $0.00557
Sonnet 5.5 $0.00009 $0.00278
Haiku 4.5 $0.00005 $0.00139

Measured 1mo ago against content hash 13e558ef7107, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, 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 1mo ago.

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 · 95 lines

How it starts

The opening of the file, as written. The whole thing — 95 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Handoff

The user is about to clear the context window. Produce a prompt that lets a fresh session, which starts with zero memory of this one, continue the work without re-deriving anything this session established.

Today's date, time and time-zone offset: !date +%FT%H:%M%z Current branch and HEAD: !git branch --show-current 2>/dev/null | grep . || echo none at !git rev-parse --short HEAD 2>/dev/null || echo none Uncommitted files: !git status --short 2>/dev/null | wc -l | tr -d ' ' Argument given by the user (may be empty): $ARGUMENTS

Procedure

  1. Decide the slug. If the user gave an argument, use it (lower-case, hyphens). Otherwise derive a short slug from the next work item (the row number or feature name the next session will work on). The file is tmp/docs/<today>-handoff-<slug>.md, where <today> is the date part (YYYY-MM-DD) of the date and time printed above. Create tmp/docs/ if it does not exist. Never overwrite an existing file: if the name is taken, append -2.

  2. Collect the facts, mostly from files, not from memory. Every source below is optional; use the ones the project has. Read state.md and the last two [saved] entries of session-log.md (written by the context-management skill in projects where state was saved); the previous handoff in tmp/docs/ (the newest *handoff* file) for the conventions section, or the project's CLAUDE.md when there is no previous handoff; the version file and the installed version if the project publishes a package or plugin; the project's worklist or issue file if state.md, the previous handoff or CLAUDE.md names one (the item the next session works on). Delegate bulk reads to a subagent; keep this window small.

  3. Write the prompt with exactly these sections, in plain English, no idioms, every technical term defined at first use, lines at most 88 columns:

    • Header line. The file's first line is exactly Handoff: written=<YYYY-MM-DD>T<HH:MM><+hhmm> branch=<name|none> head=<short sha|none>, with the values printed above (<+hhmm> is the time-zone offset, such as +0200; <short sha> is the short commit id; none for an empty value). The /pickup skill reads it to find the newest handoff and the commits made after it. When the next task has an end state that a commit may not show, the second line is Done when: <one checkable condition>.
    • Opening paragraph. The task in one or two sentences, naming the row or feature and the file and line where it is described; the repository, the branch, the version, whether the tree is clean, and whether the installed plugin matches main.
    • ## What is already known (do not re-derive). Every fact this session verified that the next session would otherwise have to re-measure: figures, with the transcript ids or file:line they came from; decisions the user made, stated as decisions; contradictions found between documents and evidence; what was rejected and why, in one line each; the exact commands that produced key evidence when they were not obvious. Facts only, each with its source.
    • ## How to proceed. Numbered steps, in the order the next session should take them. Name the design step when one is needed (independent lenses plus a rebuttal round before tallying), the point where the user decides, and the delivery shape (spec plus orchestrated run, or a direct change on a feature branch with assertions first). Name every fast test suite to run.
    • ## Conventions that bit us. The repository rules that cost time in this or earlier sessions: forbidden words, line budgets, merge style, the full release checklist with every file to bump, the issues-log rules, the save-state rule. Carry forward the previous handoff's list and add what this session learned; drop nothing unless it is no longer true.
    • Closing reminders after the sections: uncommitted files, background agents still running, a CLI restart owed after a plugin update, files that are local and untracked.

Read the full file on GitHub · 95 lines

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. 1mo ago First seen · 95 lines · 46 tokens per session scan A 13e558ef7107

Subscribe to this mod's changes

handoff is a skill published in the GitHub repository brunob54/superpowers-orchestrator (3 stars, last pushed yesterday), licensed MIT. It adds 46 tokens to every session and 1,392 once invoked, about $0.0002 per session on Opus 5.5. 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 2026-09-16.