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 branch-recapgit 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/branch-recap)<a href="https://agentmods.dev/skills/malo-t/claude-toolbelt/branch-recap"><img src="https://agentmods.dev/badge/skills/malo-t/claude-toolbelt/branch-recap/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/branch-recap"><img src="https://agentmods.dev/badge/skills/malo-t/claude-toolbelt/branch-recap.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 | $0.00338 | $0.07332 |
| Opus 5 | $0.00169 | $0.03666 |
| Sonnet 5 | $0.00068 | $0.01466 |
| Haiku 4.5 | $0.00034 | $0.00733 |
Grade A, and why
branch-recap 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 11d 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.
How it starts
The opening of the file, as written. The whole thing — 578 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Branch Recap
Tell the user what happened on their branch, so they can settle it before pushing. The situation this exists for: the work got produced fast — often with an assistant holding the keyboard — the result works, and nobody really knows what it changed. Taking stock is the job.
So the deliverable is what was done, why, how, and what deserves a second look, group by
group, and then one question set to steer from. Not the code: the user has git diff, and
what they lack is the account of it. Quote a few lines only where a remark is unreadable
without them, and let them ask for any diff they want to see.
Two failure modes to avoid. Reprinting the branch: a wall of hunks is unreadable in a terminal, and every line of it pushes the decisions further down the scroll. And chopping the account into prompts: a question between every two groups scatters the remarks across a corridor of round-trips. One account, one question set, one round of fixes.
Invocation
/branch-recap [optional restriction]
Examples:
/branch-recap/branch-recap les 2 derniers commits/branch-recap seulement le back- "je voudrais repasser sur ce que j'ai fait avant d'ouvrir la MR"
- "on a produit ça en deux heures, qu'est-ce qui a changé au juste avant que je pousse ?"
Step 1 — Resolve the scope
Never start reading before you know exactly what you are reading and the user has seen it.
The default is the current branch against its parent, up to their last common commit. Everything else is the user's to name — "les 3 derniers commits", "ce qui touche au composant paiement", "juste ce que je n'ai pas poussé". Take what they said; only work out the default when they said nothing.
Git has no notion of a parent branch, so it has to be inferred: among the integration
branches that exist — origin/HEAD, origin/develop, origin/main, origin/master — take
the one whose merge-base with HEAD is closest, in commits. That winner is <base>, the
single name used for it everywhere below:
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.
- 11d ago First seen · 578 lines · 338 tokens per session scan A ad01de875ffb
branch-recap is a skill published in the GitHub repository Malo-T/claude-toolbelt (2 stars, last pushed 21d ago), licensed MIT. It adds 338 tokens to every session and 7,332 once invoked, about $0.0017 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.
Other skills, from other repositories
github-audit
Audit GitHub repo health, auto-fix safe issues, run quality gates, deploy, and record decisions across runs. Use when asked to audit a repo, check repo/CI/dependency health, run pre-deploy quality checks, or ship after an audit.
commit-push-pr
Full git workflow - creates branch, commits, pushes, and creates or updates a PR with summary and test plan.
commit
Create a single well-crafted git commit from current changes. Analyzes diff, follows repo's commit style, and writes a concise "why not what" message.
update-architecture-atlas
Create or refresh a versioned Curie architecture-atlas JSON snapshot from a git diff, including current and vision flows, maturity-rated seams, ADRs, and documentation drift. Use when asked to update, regenerate, compare, or add a version to the interactive architecture atlas.
release
Automated release management - version updates, changelog generation, git tagging, and GitHub release creation.
pr-conventions
Author PR descriptions from the right per-change-type template (the repo's own PULLREQUESTTEMPLATE wins), embed issue-tracker links, keep the description current as fixes land, and reply to human and bot review threads. INVOKE when opening or updating a PR or responding to reviewer comments — inside /sonu:ship or…