backup

backup is a skill for Claude Code from mishahanin/heading-os. It costs 24 tokens per session (1,631 once invoked), scanned A, original, Apache-2.0.

A workspace backup tool that commits and pushes both the code repository and its separate data repository to GitHub.

In plain words
What is it for?
Use it to preview, commit, and push workspace changes, or push existing commits without creating a new commit.
Why use it?
It avoids forgetting the data repository or leaving changes unsaved when backing up manually. It also checks for tracked secrets before pushing.

Skill for Claude Code

Written for Claude Code: allowed-tools in frontmatter. Also seen: model in frontmatter; reads .claude/ paths.

Needs its repository: it runs a file that does not travel with it, so clone the repository first. The line is python scripts/push-all.py # commit working-tree changes + push both.

Good fit Use it to preview, commit, and push workspace changes, or push existing commits without creating a new commit.

Compare 6 skills from other repositories ↓
Install

Getting it into your agent

It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.

Clone the repo
git clone --depth 1 https://github.com/mishahanin/heading-os
agentmods
npx agentmods add skills/mishahanin/heading-os/backup

Made for: Claude Code.

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 backup

README.md
[![agentmods](https://agentmods.dev/badge/skills/mishahanin/heading-os/backup.svg)](https://agentmods.dev/skills/mishahanin/heading-os/backup)
Your own site
<a href="https://agentmods.dev/skills/mishahanin/heading-os/backup"><img src="https://agentmods.dev/badge/skills/mishahanin/heading-os/backup.svg" alt="Measured on agentmods" height="20"></a>
Per session 24 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 1,631 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. Third-party audits
  • NVIDIA SkillSpector warn 7 Sept 2026
SkillSpector: 1 finding, up to medium

These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →

  • medium Excessive Agency · line 11
    Skill selects an external model or provider that may use a different account or billing plan than the operator expects. Undisclosed model switches can cause unexpected cost or quota consumption.
    Fix: Remove the model/provider override or disclose it prominently and require explicit operator approval before invoking an external coding CLI or billed model.
How audits are shown
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 $0.00024 $0.01631
Opus 5 $0.00012 $0.00816
Sonnet 5 $0.00005 $0.00326
Haiku 4.5 $0.00002 $0.00163

Measured 7d ago against content hash 2e6bedd15720, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-07, from the pricing page.

Security

Grade A, and why

backup 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 7d 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.

.claude/skills/backup/SKILL.md · 140 lines

How it starts

The opening of the file, as written. The whole thing — 140 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Backup the entire workspace to GitHub.

CEO two-part topology (engine + data): use push-all.py

The CEO workspace is split into two repos, each with its own private origin/main. The ENGINE clone (.heading-os) holds code only. The DATA overlay (.heading-os-data) holds all data and every artifact. One command commits and pushes BOTH. It runs a pre-push secret scan and an ahead/behind [0 0] verification, because a bare push can silently leave a ref behind:

python scripts/push-all.py            # commit working-tree changes + push both
python scripts/push-all.py -m "msg"   # custom commit message
python scripts/push-all.py --no-commit  # push existing commits only
python scripts/push-all.py --dry-run    # show what would happen, change nothing

It reads GH_TOKEN from the engine .env, refuses to push any tracked secret-like file (.env, .session, cookies.json, .sessions/), and never pushes .memory-index/ (gitignored, rebuildable). The DATA overlay is pushed FIRST. The engine's pre-push hook runs the full suite inside the push, and data is the only half that cannot be reconstructed. Prefer this over the manual git steps below whenever the data overlay exists (get_data_root() differs from the engine root). The manual steps remain the path for exec workspaces and the pre-cutover single-repo case.

Exit 3 means at least one repository was skipped. Read the HEADLINE, not the code, because it has two shapes and they mean opposite things to the operator.

  • Partial: N of M repo(s) not pushed. — the backup was partial, not failed. Report it that way. The repositories named below the headline were not pushed, for the reason printed next to each; everything else was pushed and verified.
  • NOTHING PUSHED: all M repo(s) skipped. — nothing reached a remote and there is NO new off-machine copy. Never report this as a partial success. The exec and pre-cutover single-repo modes push one repository, so exit 3 there is always this shape.

In both shapes every skipped repository is still committed locally, so nothing is lost. Tell the operator which repository was skipped, quote the reason, and give the remedy. The remedy is usually to merge the branch into main and run the command again. For an unarmed engine test gate, it is python scripts/install-git-hooks.py. Do NOT re-run the command hoping for a different answer.

Exit 1 and 2 are real failures that stopped the run. The causes are a security refusal, a remote a repository must not push to, or an absent GH_TOKEN. They are also a misconfigured data root, or a push that ran and did not verify.

Exit 2 with "REFUSING TO PUSH" naming a remote is the newest of those and is different in kind from the others. It means a repository is aimed at a remote it must not push to: the engine's own remote, or a repository GitHub reports as public. Nothing was pushed and nothing was lost. The whole run stops rather than continuing with the other repositories, because a wrong remote makes all of them suspect. Report the remote it names to Misha and do not retry until it is corrected.

Because the DATA overlay is pushed first, a failure at the ENGINE can still leave DATA pushed and verified. Read the per-repository lines above the failure before you tell the operator nothing was backed up. Exit 0 means everything went.

Read the full file on GitHub · 140 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. 7d ago First seen · 140 lines · 24 tokens per session scan A 2e6bedd15720

Subscribe to this mod's changes

backup is a skill published in the GitHub repository mishahanin/heading-os (11 stars, last pushed yesterday), licensed Apache-2.0. It adds 24 tokens to every session and 1,631 once invoked, about $0.0001 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 2026-08-30.