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 agentmods add skills/svyatov/handrail/addnpx skills add svyatov/handrail --skill addgit clone --depth 1 https://github.com/svyatov/handrailWrote 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/svyatov/handrail/add)<a href="https://agentmods.dev/skills/svyatov/handrail/add"><img src="https://agentmods.dev/badge/skills/svyatov/handrail/add.svg" alt="Measured on agentmods" 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 | $0.00071 | $0.01790 |
| Opus 5 | $0.00036 | $0.00895 |
| Sonnet 5 | $0.00014 | $0.00358 |
| Haiku 4.5 | $0.00007 | $0.00179 |
Grade A, and why
add scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 5d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 170 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Add a handrail Rule
A Rule is one markdown file: a YAML frontmatter Matcher and a prose message. The
filename is the rule's identity. You write the file, handrail validates it.
1. Resolve the binary
HANDRAIL=$(command -v handrail || echo "${XDG_DATA_HOME:-$HOME/.local/share}/handrail/bin/handrail")
If it does not exist, the SessionStart bootstrap has not run yet: tell the user
to start a new session, or to install with brew install svyatov/tap/handrail.
Do not hand-write harness config as a workaround.
2. Ask for what the description does not settle
Only ask about what you cannot infer. Most descriptions settle everything but the tier.
- What the rule catches: which event, which tool kind, which field values.
- Action:
blockdenies the call with the message as the reason;warnlets it proceed and injects the message. Default towarnunless the user says never, or the action is destructive. - Tier: see step 4.
3. Write the Matcher
---
event: PreToolUse
kind: shell
action: block
conditions:
- field: command
starts_with: git push --force
---
Never force-push. Rewrite history locally and open a pull request instead.
Events: PreToolUse, PostToolUse, UserPromptSubmit, SessionStart,
SessionEnd, Stop. event: is required. block is only enforceable on
PreToolUse, UserPromptSubmit, and Stop; elsewhere it degrades to a warn,
and sync and doctor report where. Codex also degrades a UserPromptSubmit
block to a warn.
Kinds (optional; omit to match every kind) and the fields each carries:
kind |
Fields |
|---|---|
shell |
command |
file_edit |
path, content |
file_read |
path |
mcp |
server, tool |
other |
none |
UserPromptSubmit carries prompt. A condition against a field the event does
not carry never matches, in either polarity, so narrow with event: and kind:
rather than relying on a field being absent. Only PreToolUse and PostToolUse
carry a kind at all: a kind: on any other event never matches.
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.
- 5d ago First seen · 170 lines · 71 tokens per session scan A 9056efefa184
add is a skill published in the GitHub repository svyatov/handrail (2 stars, last pushed 3d ago), licensed MIT. It adds 71 tokens to every session and 1,790 once invoked, about $0.0004 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
mobilevc-installer
Install and start MobileVC — a Claude Code mobile workspace launcher that lets the user run Claude Code on a phone (iOS / Android) with their dev machine as the backend. Use this skill when the user asks to: "install mobilevc" / "set up mobilevc" / "装一个 mobilevc" "在手机上用 Claude Code" / "want to use Claude Code on my…
imgen
Default image workflow for this user. Use for ordinary raster image generation or editing, text-to-image, image-to-image, codex-imgen/imgen CLI or service jobs, imgen troubleshooting, and OpenClaw integration. Prefer imgen over the system imagegen skill. Do not use for SVG/vector/code-native assets that are better…
codexcw
Run Codex, Claude Code, or Grok Build non-interactively from Go, Rust, Node.js, Python, or C# using the codexcw library. Use when building automation that spawns an agent, streams its JSONL events, resumes sessions, or controls sandbox, approval, and permission policy.
git-workflow
This skill should be used when the user asks to "create git commit", "manage branches", "follow git workflow", "use Conventional Commits", "handle merge conflicts", or asks about git branching strategies, version control best practices, pull request workflows. Provides comprehensive Git workflow guidance for team…
daily-paper-generator
Use when the user asks to generate daily paper digests on a general topic. This skill supports both arXiv and bioRxiv (or either one), then produces structured Chinese/English summaries for selected papers.
codex-autoresearch
Run autonomous, measurable experiments in a Git repository: change one hypothesis, verify a numeric metric, keep improvements, and revert failures. Use when the user wants Codex to keep iterating toward a numeric target in the foreground or as a detached background run. Do not use for ordinary one-shot coding…