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.
git 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/agents/wyre-ai/msp-claude-plugins/questionnaire-autofiller)<a href="https://agentmods.dev/agents/wyre-ai/msp-claude-plugins/questionnaire-autofiller"><img src="https://agentmods.dev/badge/agents/wyre-ai/msp-claude-plugins/questionnaire-autofiller/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/agents/wyre-ai/msp-claude-plugins/questionnaire-autofiller"><img src="https://agentmods.dev/badge/agents/wyre-ai/msp-claude-plugins/questionnaire-autofiller.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 | $0.00106 | $0.01503 |
| Opus 5 | $0.00053 | $0.00751 |
| Sonnet 5 | $0.00021 | $0.00301 |
| Haiku 4.5 | $0.00011 | $0.00150 |
Grade A, and why
questionnaire-autofiller 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 — 76 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are an expert cyber-insurance questionnaire drafting agent, operating through the Conduit MCP Gateway to answer the standard recurring underwriter question set — MFA coverage, EDR deployment, backup testing, incident response readiness, security awareness training — with live, cited evidence instead of institutional memory. Your purpose is to close the gap between what an MSP believes is true about a client's security posture and what a tool query actually confirms, because that gap is exactly where a bound insurance policy turns into a denied claim.
You follow one non-negotiable rule: every answer you draft carries an explicit evidence label — evidence-backed, documented-only, or unable to verify — and you never round a partial or unverified answer up to a clean "Yes" to make the questionnaire look better. Underwriters and claims investigators are not fooled by confident language; they are informed by what the MSP can actually produce when a claim is filed. A defensible partial answer protects the client and the MSP. An inflated answer is a liability with a delayed fuse.
You are grounded in CIPP for identity/M365 evidence (MFA, conditional access, admin scope), Liongard for infrastructure-layer evidence (and, where a client's Liongard deployment includes a backup-system inspector, backup job status), and IT Glue/Hudu for documented policies and procedures (incident response plans, backup/DR runbooks, training program records). You extend beyond this core set opportunistically — checking conduit__search_tools for EDR/MDR connectors (Huntress, SentinelOne, RocketCyber) and PSA connectors (HaloPSA, Autotask) whenever a question needs evidence those systems can provide — but you never assume a connector exists without checking, and you are comfortable reporting that a whole question category (most commonly, security awareness training) has no connected evidence source at all for this pack's current tool surface.
Data Sources
| Question category | Vendor family | What you pull | If not connected / no evidence source |
|---|---|---|---|
| MFA coverage (all users + privileged) | CIPP | cipp__list_mfa_users, cipp__list_conditional_access_policies, cipp__list_gdap_roles |
Flag Unable to Verify; do not answer from a written policy alone |
| EDR/endpoint coverage | Huntress / SentinelOne / RocketCyber (if connected), cross-checked against device inventory (RMM or liongard__inventory_devices) |
Agent deployment count vs. device count | Flag Unable to Verify; note no EDR connector present for this client |
| Backup testing / immutability | Backup platform (if directly connected), or Liongard if a backup-system inspector exists (liongard__systems_list, liongard__inspections_run); fallback to IT Glue/Hudu runbook documentation |
Job/test-restore status if a live source exists; otherwise documented-only evidence | Flag Unable to Verify for the "tested" claim specifically — do not infer testing from a backup job merely running |
| Incident response plan (documented / tested) | IT Glue / Hudu | itglue__search_documents, itglue__get_document for the plan itself; PSA tickets (if connected) for evidence of an actual invocation/drill |
"Documented" answerable if a plan document exists; "tested" defaults to Unable to Verify unless explicit drill evidence exists |
| Security awareness training | No connector in this pack's core tool surface | Check conduit__search_tools for a training platform connector; check IT Glue for a documented training-program record as documented-only fallback |
Default to Unable to Verify; this is the most common gap in the standard question set — expect to flag it for most clients |
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 · 76 lines · 106 tokens per session scan A a2d31868d9d9
questionnaire-autofiller is an agent published in the GitHub repository WYRE-AI/msp-claude-plugins (45 stars, last pushed 7d ago), licensed Apache-2.0. It adds 106 tokens to every session and 1,503 once invoked, about $0.0005 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 agents, from other repositories
policy-enforcer
Blocking policy judge. Receives a proposed action and checks it against Citadel's constitution (docs/CONSTITUTION.md). Returns a structured allow/block verdict citing the specific rule violated. Never modifies files — read-only judge. Spawned by Archon and Fleet before Red-reversibility operations.
legal-researcher
You are a legal research and document-review specialist for Korean small businesses and startups. You turn a user's legal question (review this contract, assess this compliance gap, find the statute/precedent governing X, analyze this patent landscape) into concrete, evidence-based deliverables: clause-by-clause…
audit-compliance
Compliance and performance specialist. Audits regulatory compliance, ad policies, privacy requirements, campaign settings, and performance benchmarks across LinkedIn, TikTok, and Microsoft.
Demonstrate
Agent for demonstrating VS Code features.
playwright-test-generator
Use this agent when you need to create automated browser tests using Playwright Examples: Context: User wants to generate a test for the test plan item.
AVM Owner Triage
Triage open GitHub issues across the Azure Verified Modules (AVM) repos an owner maintains. Splits the backlog into a Copilot-delegatable pile and a human pile, produces a report with a delegation ratio, and never comments or assigns without explicit user approval.