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.
npx skills add Malo-T/claude-toolbelt --skill handoffgit clone --depth 1 https://github.com/Malo-T/claude-toolbeltWrote 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.
[](https://agentmods.dev/skills/malo-t/claude-toolbelt/handoff)<a href="https://agentmods.dev/skills/malo-t/claude-toolbelt/handoff"><img src="https://agentmods.dev/badge/skills/malo-t/claude-toolbelt/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.
<a href="https://agentmods.dev/skills/malo-t/claude-toolbelt/handoff"><img src="https://agentmods.dev/badge/skills/malo-t/claude-toolbelt/handoff.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00151 | $0.00691 |
| Opus 5.5 | $0.00060 | $0.00276 |
| Sonnet 5.5 | $0.00030 | $0.00138 |
| Haiku 4.5 | $0.00015 | $0.00069 |
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.
How it starts
The opening of the file, as written. The whole thing — 58 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Session handoff
Two modes. Write is the default; resume applies when the user gives a path or asks to resume.
Write mode
1. Where
${CLAUDE_CONFIG_DIR:-~/.claude}/handoffs/<repo-or-directory-name>/<YYYY-MM-DD-HHmm>-<short-slug>.md
Outside the repo, so git stays clean. For a handoff the user will pick up later ("tomorrow", "later"), use the suffix .later.md: the plugin then leaves the session alone, so skip step 3. If unsure which variant is wanted, ask before writing.
2. Content
The reader is a fresh Claude with no context. Anything unwritten is lost. Be factual and dense, not narrative.
# Handoff: <title>
Written <date and time> from <cwd>, branch <branch>.
## Goal
<what the session set out to do, and why>
## Where things stand
<what is done, verified, and how it was verified>
## Findings and rejected hypotheses
<what was tried or ruled out, so it is not redone>
## Decisions
<each decision with its reason>
## Uncommitted changes
<every modified file in the tree, one line each, with what changed>
## Next step
<the single next action, then what follows>
## Open questions
<what only the developer can answer>
If the user wants only part of the work carried over, put the scope at the top, develop only that part, and list the rest under "Out of scope", one line per topic. "Uncommitted changes" still lists every modified file, in scope or not, so nothing is lost or committed by mistake.
3. Hand over to the plugin
Write nothing else and run nothing. The plugin's mod notices the file written under handoffs/, and once your turn ends it runs /clear and submits the resume prompt in the fresh session. End the turn with one line naming the file; do not tell the user to type /clear.
Resume mode
Without a path, take the most recent file in ${CLAUDE_CONFIG_DIR:-~/.claude}/handoffs/<repo-or-directory-name>/ and name it before continuing. Read it, say in three lines where things stand, check that the described state is still true (git status, the files it cites), and propose the next step. Run nothing before the user agrees. For a partial resume ("only X"), work on that part and leave the others untouched.
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.
- yesterday First seen · 58 lines · 151 tokens per session scan A 3f57116829ed
handoff is a skill published in the GitHub repository Malo-T/claude-toolbelt (2 stars, last pushed yesterday), licensed MIT. It adds 151 tokens to every session and 691 once invoked, about $0.0006 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 2026-11-12.
handoff compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| handoffMalo-T/claude-toolbelt | Skill | 2 | yesterday | MIT | A | 151 | Claude Code |
| magic-compactaerovato/magic-compact | Skill | 176 | 1mo ago | BSD-3-Clause | A | 10 | Claude Code |
| reflectsneg55/agent-starter | Skill | 75 | 2mo ago | MIT | A | 37 | Claude Code |
| remembersneg55/agent-starter | Skill | 75 | 2mo ago | MIT | A | 38 | Claude Code |
| dreamsneg55/agent-starter | Skill | 75 | 2mo ago | MIT | B | 27 | Claude Code |
| cold-outreachadamjali/claude-skills | Skill | 1 | 8d ago | MIT | A | 118 | Claude Code |
| email-broadcastsadamjali/claude-skills | Skill | 1 | 8d ago | MIT | A | 102 | Claude Code |