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 WYRE-AI/msp-claude-plugins --skill restore-test-verificationgit clone --depth 1 https://github.com/WYRE-AI/msp-claude-pluginsWrote 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/wyre-ai/msp-claude-plugins/restore-test-verification)<a href="https://agentmods.dev/skills/wyre-ai/msp-claude-plugins/restore-test-verification"><img src="https://agentmods.dev/badge/skills/wyre-ai/msp-claude-plugins/restore-test-verification/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/wyre-ai/msp-claude-plugins/restore-test-verification"><img src="https://agentmods.dev/badge/skills/wyre-ai/msp-claude-plugins/restore-test-verification.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
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 | $0.00077 | $0.01800 |
| Opus 5 | $0.00039 | $0.00900 |
| Sonnet 5 | $0.00015 | $0.00360 |
| Haiku 4.5 | $0.00008 | $0.00180 |
Grade A, and why
Restore-Test Verification 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 6d 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 — 157 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Restore-Test Verification
Overview
A backup job reporting "success" proves that data was copied somewhere. It proves nothing about whether that data can be turned back into a working system or a usable file when someone actually needs it. Corruption, encryption-key problems, application- consistency failures, and format incompatibilities routinely surface only at restore time — never during the backup run itself. The gap between "backed up" and "recoverable" is where most DR failures actually happen, and it is invisible from a job-success dashboard. This skill exists to make that gap visible: tracking what restore-test evidence actually exists per protected system, and flagging where it doesn't.
This skill assumes backup jobs are already running (see backup-job-health for that
layer) and asks the next question: if this needed to be restored today, is there any
evidence it actually could be?
Anti-triggers
- Fetching a screenshot or boot-verification record — the screenshot,
activity, and asset endpoints are the appliance's own API surface; use
datto-bcdr-api-patterns. This skill weighs what that evidence proves and flags where none exists. - Answering "are backups tested?" on an insurance or audit form — the
evidence-labelling discipline for that context lives in
insurance-questionnaires(compliance-pack), which draws on this skill's finding rather than reproducing it.
Key Concepts
"Exists" vs. "recoverable" evidence, ranked
Not all restore evidence is equally strong. From strongest to weakest:
- Actual restore performed — a real recovery event (production failover, file recovery request, or a deliberate DR drill) where the data was genuinely restored and validated. Strongest possible evidence; also the rarest.
- Full boot/virtualization verification — the protected system was virtualized from its backup image and confirmed to boot and respond (some BCDR appliances, e.g. Datto BCDR, run this automatically). Strong evidence for image-based appliance backups — it proves the image is bootable, though it doesn't fully validate application-level data integrity inside the guest.
- Screenshot verification — an automated screenshot of the virtualized boot
screen, confirming the OS reached a login prompt or desktop rather than a boot
error. Weaker than full boot verification but far stronger than nothing — it's
the closest-to-continuous proxy most appliance vendors offer. Treat a "successful
backup" with a failed screenshot verification as a recovery risk, not a pass —
see
backup-job-healthfor the point that both signals must be checked together. - Spot-check restore drill — a manual, periodic restore of a sample file, mailbox, or object to confirm data integrity, without a full system failover. This is the primary evidence type available for SaaS-data snapshot backups (Datto SaaS Protection, Spanning), which generally don't offer a "boot" concept at all — a spot-check restore of a mailbox or SharePoint file is the equivalent verification action.
- No test evidence at all — a backup with a clean job-success history and no restore, boot-verification, screenshot, or spot-check record of any kind. Treat this explicitly as untested, not merely "unverified" — it should be reported as a finding in its own right, not omitted because nothing failed.
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.
- 6d ago First seen · 157 lines · 77 tokens per session scan A 454eef1c97d1
Restore-Test Verification is a skill published in the GitHub repository WYRE-AI/msp-claude-plugins (45 stars, last pushed 8d ago), licensed Apache-2.0. It adds 77 tokens to every session and 1,800 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 2026-09-04.
Other skills, from other repositories
live-preview
Mid-build visual verification loop. Takes screenshots of components during construction, not just after. Catches visual regressions and invisible features before they compound. Requires Playwright or similar screenshot tool.
content-card-news
A planning tool for four-card social media carousels on Instagram, Threads, and Kakao Channel, combining card copy, design guidance, image prompts, captions, and hashtags.
cs-voc-triage
A customer-feedback analysis tool that combines reviews and support contacts from multiple channels. VOC means “voice of the customer”: the comments, questions, complaints, and praise customers leave about a business.
design-copywriting
A copywriting workflow for marketing and product pages that uses a brand voice and returns structured text for page sections.
data-realestate
A lookup tool for reported South Korean property transactions using Ministry of Land data. It covers sales and rental deals for apartments, officetels, multi-family homes, houses, and commercial or office properties.
meta-feedback
A feedback reporter that collects bug reports and feature requests, then formats them as GitHub Issues. GitHub Issues are shared records used by development teams to track work and problems.