Borrowing it
Nothing to install: this file belongs to webteractive/zetty. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/webteractive/zetty/main/.claude/skills/verify/SKILL.mdgit clone --depth 1 https://github.com/webteractive/zettyWrote 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/webteractive/zetty/verify)<a href="https://agentmods.dev/skills/webteractive/zetty/verify"><img src="https://agentmods.dev/badge/skills/webteractive/zetty/verify/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/webteractive/zetty/verify"><img src="https://agentmods.dev/badge/skills/webteractive/zetty/verify.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 | $0.00046 | $0.00517 |
| Opus 5.5 | $0.00018 | $0.00207 |
| Sonnet 5.5 | $0.00009 | $0.00103 |
| Haiku 4.5 | $0.00005 | $0.00052 |
Grade A, and why
verify 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 12d 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.
What it actually says
Verifying Zetty
The recipe is .hydra/rules/isolated-test-app.md: build, freeze a copy in
/tmp/zti, launch it as a background task with CFFIXED_USER_HOME and
ZMX_DIR, drive it with its own binary (iz), tear down with
iz quit --kill-sessions --force, and check the real zmx list count is
unchanged. Never quit or replace /Applications/zetty.app to verify.
Reading what happened
- Pane screen:
ZMX_DIR=/tmp/zti/zmx env -u ZMX_SESSION zmx history zetty-<id>; add--vtfor styling (SGR 2 is faint, which is what Claude's placeholders are drawn in). - App log:
/usr/bin/log show --last 2m --predicate 'subsystem == "co.webteractive.zetty"' --style compact(not zsh'slog). Hibernation logshibernate:andhandoff:lines. iz statusshows‹handoff: writing|ready|failed›.
Hibernate and handoffs
- Each Claude test goes in a FRESH directory: without hooks the pane's session is guessed from the newest transcript there.
- A new directory first shows Claude's trust question:
iz send --pane <id> --key Down --key Enter. iz hibernate <project> --force(an agent pane counts as busy).- Claude's predicted-prompt placeholder needs
CLAUDE_CODE_ENABLE_PROMPT_SUGGESTION=1on theclaudecommand line; a pane woken from a handoff resumes without it. - The queued-message hint shows while a long reply runs ("Count from 1 to 300…") and a second message is sent behind it.
- Without hooks a working Claude is not known to be working, so a hibernate
mid-turn types
/compactover the turn. That path needs the installed app. - Delete the throwaway chats afterwards:
~/.claude/projects/-private-tmp-zti-*folders the test made.
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.
- 12d ago First seen · 38 lines · 46 tokens per session scan A c3c37be3bff6
verify is a skill published in the GitHub repository webteractive/zetty (2 stars, last pushed yesterday), licensed MIT. It adds 46 tokens to every session and 517 once invoked, about $0.0002 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 2026-10-18.
Other skills, from other repositories
inspect-appkit-ui
Inspect and review a live native macOS AppKit interface in Codex Browser. Use when the user asks Codex to preview a running AppKit app, click or identify an NSView, map a visual element to its class and hierarchy, diagnose layout or accessibility issues, or prepare precise visual feedback for Codex.
wispterm-diagnostics
Use when a user wants to report, troubleshoot, or collect context for a WispTerm issue, including crashes, rendering/DPI glitches, high CPU, keyboard/input bugs, selection/copy/scrolling, SSH/SCP failures, SSH image preview failures, HTML preview/browser panel failures, SSH disconnects such as…
cmux-test-bisect
Find when and why tests broke on main: bisect a Swift package suite on CI across old commits, read the failure matrix, and decide per test whether the test went stale or the code regressed. Use when a suite is red on main, when PR CI only ran filtered tests and the full suite has drifted, or when asked which PR broke…
roundtrip-radar
Per-journey code audit tracing data through complete user flows for bugs, data safety, performance, and round-trip completeness. Discovers workflows, audits each end-to-end, rolls up cross-cutting issues, and supports natural-language flow tracing. Triggers: "roundtrip audit", "trace user journey", "/roundtrip-radar".
liney-cli
Use the Liney CLI (liney) to inspect or control a running Liney terminal workspace and the agent sessions it hosts. Liney runs several coding agents in parallel, each in its own pane/tab/worktree, so reach for this whenever the user wants to act on a pane other than the current one — check on, coordinate, read from…
macos-test-diagnoser
Before invoking Apple-only binaries, confirm the execution context is macOS. From Windows or Linux, run those steps in a Mac SSH project or through an already configured remote transport; do not retry missing Apple binaries locally.