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 Tx1207/hello-scholar --skill remote-rungit clone --depth 1 https://github.com/Tx1207/hello-scholarWrote 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/tx1207/hello-scholar/remote-run)<a href="https://agentmods.dev/skills/tx1207/hello-scholar/remote-run"><img src="https://agentmods.dev/badge/skills/tx1207/hello-scholar/remote-run/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/tx1207/hello-scholar/remote-run"><img src="https://agentmods.dev/badge/skills/tx1207/hello-scholar/remote-run.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 | $0.00086 | $0.01039 |
| Opus 5.5 | $0.00034 | $0.00416 |
| Sonnet 5.5 | $0.00017 | $0.00208 |
| Haiku 4.5 | $0.00009 | $0.00104 |
Grade A, and why
remote-run 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 4d 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 — 45 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Remote Run
One agent works on the local copy; another, such as Codex, edits, runs, and debugs on the server copy, next to the GPUs and data. Both copies are Git work trees on the same branch. Changes move between them only as commits, through scripts/remote/sync.sh on the local machine, and only by fast-forward. Data, weights, logs, and outputs stay untracked on the server, so sync never touches them.
Confirm the facts first
Ask for the SSH alias and the project path on the server. Check key login with ssh -o BatchMode=yes <alias> hostname, Git on the server, and that the local project is a Git repository with a branch checked out. Do not edit ~/.ssh/config, install packages, or change server settings yourself; when login needs a key or connection reuse, show the user the config block to add, including ControlMaster auto, ControlPersist, and ServerAliveInterval. Setting up the server agent is the user's job.
Survey both sides read-only: the server's top-level entries with their sizes, and the local .gitignore. Directories for data, checkpoints, and outputs, such as data/, artifacts/, outputs/, or wandb/, must be ignored by Git before the first sync. Recommend the .gitignore lines as a table, add them after the user confirms, and follow the layout recorded in Architecture.
Install the script
Copy the files in assets/scripts/remote/ into the project's scripts/remote/, keep sync.sh executable, fill in remote.env, and commit them.
| Command | Effect |
|---|---|
sync.sh init |
First time: make REMOTE_DIR a work tree on the local branch without overwriting server files. |
sync.sh status |
Commits each side lacks, uncommitted server files, and the server's tmux jobs. |
sync.sh pull |
Fast-forward the local branch to the server's commits. |
sync.sh push |
Fast-forward the server branch to the local commits; needs a committed local tree. |
pull and push stop without changing either side when the server is on another branch, when it has uncommitted changes to tracked files, or when both sides have commits the other lacks; a push also stops when it would overwrite an untracked server file. The project's rules tell the local agent what to do next: ask for a commit on the server, or rebase its local commits onto refs/remote-run/server and push. Untracked server files that Git does not ignore are listed, never pulled.
What ships with it
6 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 4d ago First seen · 45 lines · 86 tokens per session scan A a7c6ca1745f4
remote-run is a skill published in the GitHub repository Tx1207/hello-scholar (10 stars, last pushed 4d ago), licensed MIT. It adds 86 tokens to every session and 1,039 once invoked, about $0.0003 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 2026-10-26.
Other skills, from other repositories
work
Handle issue/PR work items with worktree isolation. Triggered by "/work [issuelink|description]" to start focused work.
atmos-git
Atmos Git and GitOps: git.repositories, clone/pull/status/diff/commit/push/clean, local Git hook shims and hook env vars, pre-commit framework migration, Starlark git.sandbox checks and unstaged fixes, signed commits, managed workdirs, fork-PR trust gate (clone gate and posting gate), and auth via identities or…
source-command-dev-loop
Fast development loop for small changes (validate → fix → commit → deploy).
source-command-quick
Fastest commit cycle - lint, fix, commit, push (10s).
sync
A guided workflow for syncing a research project with a remote repository and deployment target. It bundles common version-control commands so users do not need to know Git.
create-pr
Creates a GitHub PR with a Linear-ticket-prefixed title and a decision-led, narrative description for Prisma 8. Use when the user wants to create a pull request, open a PR, or submit changes for review.