Add google-cloud-filestore-auditing skill for fleet disaster recovery, security access governance, and architectural reliability audits of GCP Filestore instances.

PiperOrigin-RevId: 978216419
2026-09-14 20:00:20 +08:00 · 2026-09-08 18:07:33 -07:00
parent f56c173f7e
commit 4b40feb8ca
8 changed files with 1695 additions and 1 deletions
@@ -86,6 +86,7 @@ repo to install.
  - [**GKE Upgrades & Maintenance**](./skills/cloud/gke-upgrades)
  - [**GKE Workload Scaling**](./skills/cloud/gke-workload-scaling)
  - [**GKE Workload Troubleshooting Skill**](./skills/cloud/gke-workload-troubleshooting)
+  - [**Google Cloud Filestore Auditing Skill**](./skills/cloud/google-cloud-filestore-auditing)
  - [**Google Cloud Filestore Autoscale**](./skills/cloud/google-cloud-filestore-autoscale)
  - [**Google Cloud Filestore NFS File Browser**](./skills/cloud/google-cloud-filestore-nfs-browser)
  - [**Google Cloud global external Application Load Balancer Configuration Skill**](./skills/cloud/google-cloud-global-frontend-configuration)
@@ -0,0 +1,366 @@
+---
+name: google-cloud-filestore-auditing
+metadata:
+  category: Storage
+description: >-
+  Audits Google Cloud Filestore instances across projects for disaster recovery
+  readiness (missing or stale backups), security access governance (overly
+  permissive NFS export rules, 0.0.0.0/0 exposure, missing ROOT_SQUASH), and
+  reliability compliance (Physical Zone Isolation PZI and Physical Zone Separation PZS).
+  Use when assessing storage health posture, auditing NFS export permissions,
+  identifying unprotected file shares, or validating zone failure domains. Don't use
+  for Cloud Storage buckets, Persistent Disk, or NetApp Volumes.
+---
+
+# Google Cloud Filestore Auditing Skill
+
+This skill enables autonomous agents to audit, evaluate, and report the disaster
+recovery, security access governance, and architectural reliability posture of
+Google Cloud Filestore fleets across GCP projects.
+
+## Prerequisites / IAM Requirements & Permissions
+
+Before executing this skill, the runtime principal (user account or Service
+Account) must possess the following IAM roles and granular permissions on the
+target GCP project(s):
+
+### 1. Audit Operations (Read-Only Assessment)
+
+Requires the **`roles/file.viewer`** role, which provides:
+
+-   **`file.instances.list`**: Enumerate Filestore instances across project
+    locations.
+-   **`file.instances.get`**: Inspect instance configuration, NFS export rules,
+    IP ranges, squash mode, and PZI/PZS isolation status.
+-   **`file.backups.list`**: Enumerate existing backups across regions.
+-   **`file.backups.get`**: Inspect backup timestamps, source instance URIs, and
+    status.
+
+### 2. Remediation Operations (Backup Creation)
+
+Requires the **`roles/file.editor`** (or `roles/file.admin`) role, which
+provides:
+
+-   **`file.backups.create`**: Create on-demand baseline backups for unprotected
+    instances.
+-   **`file.operations.get`**: Monitor long-running backup creation operations.
+
+### 3. MCP Tool Invocation
+
+If invoking capabilities via the Google Cloud Filestore MCP Server
+(`file.googleapis.com/mcp`):
+
+-   **`roles/mcp.toolUser`**: Required to execute MCP tools (`list_instances`,
+    `get_instance`, `list_backups`, `get_backup`, `create_backup`).
+
+### 4. Cloud Authentication Setup
+
+Authenticate the runtime environment using the appropriate method for your execution context:
+
+-   **CLI Execution (`gcloud`)**:
+    ```bash
+    gcloud auth login
+    ```
+-   **Application SDKs & Scripts**:
+    ```bash
+    gcloud auth application-default login
+    ```
+-   **Automated Services & Service Accounts**:
+    ```bash
+    export GOOGLE_APPLICATION_CREDENTIALS="/path/to/service-account-key.json"
+    ```
+
+### 5. Billing Account Attachment
+
+-   **Active Billing Required**: The target GCP project must have an active Cloud Billing account attached. Filestore instances and backup operations require active billing. Verify billing status with:
+    ```bash
+    gcloud beta billing projects describe {project_id}
+    ```
+
+## Quick Start
+
+1.  Ensure `gcloud` is installed. See
+    [gcloud installation guide](https://cloud.google.com/sdk/docs/install.md.txt) if
+    needed.
+2.  Authenticate your environment:
+
+    ```bash
+    gcloud auth login
+    gcloud auth application-default login
+    ```
+3.  Set the target project context and verify an active billing account is attached:
+
+    ```bash
+    gcloud config set project {project_id}
+    gcloud beta billing projects describe {project_id}
+    ```
+4.  Configure a well-supported default region (e.g., `us-central1`). See official
+    [region availability documentation](https://cloud.google.com/about/locations.md.txt):
+
+    ```bash
+    gcloud config set compute/region us-central1
+    ```
+5.  Verify required service APIs are enabled:
+
+    ```bash
+    gcloud services enable file.googleapis.com --quiet
+    ```
+6.  Run fleet discovery and backup inventory (see "Discovery & Read Operations"
+    below).
+7.  Evaluate findings against the audit rules matrix and render the executive
+    scorecard and inventory table.
+8.  If unprotected instances are identified, present baseline backup creation
+    commands and prompt the user for confirmation.
+
+## Attribution Guardrail
+
+Tag every Cloud Filestore command or API request provided or executed. Prefix
+`gcloud` commands with the designated metrics environment:
+
+```bash
+CLOUDSDK_METRICS_ENVIRONMENT="gcs-skills gcs-skills/1.0 (skill:google-cloud-filestore-auditing)" \
+gcloud filestore instances list ...
+```
+
+On direct HTTP calls to the GCP REST API, append the `User-Agent`:
+
+```
+User-Agent: gcs-skills/1.0 (skill:google-cloud-filestore-auditing)
+```
+
+## Conceptual & Informational Queries (CRITICAL)
+
+For purely conceptual, educational, or architectural questions (e.g., *"What is
+Physical Zone Isolation (PZI) in Filestore?"*, *"Why is NO_ROOT_SQUASH
+dangerous?"*, *"Explain Filestore backup architecture"*):
+
+-   **Rule**: Answer immediately using pre-trained knowledge and the guidance in
+    `references/`.
+-   **Constraint**: **Do NOT execute external tool calls or API requests** for
+    basic conceptual queries.
+
+## Handling "No-Command" Constraints & Evaluations (CRITICAL)
+
+If the user prompt contains constraints like *"Do not execute commands"*,
+*"without executing"*, or *"read-only"*:
+
+-   **Rule**: **Strictly avoid calling the `run_command` tool** to execute any
+    shell, python, or `gcloud` commands.
+-   **Discovery Hierarchy**:
+    1.  First, check if Filestore MCP tools (`list_instances`, `list_backups`)
+        are available and query them directly (these are API invocations, not
+        shell command executions).
+    2.  If MCP tools are not present or cannot connect, search local reference
+        markdown files (specifically the mock fleet definitions in
+        `references/zone-isolation-pzi-pzs.md`) for any mock instances or
+        project details matching the request. (Do NOT attempt to read evaluation
+        config files such as `EVAL.yaml` or `EVAL.txtpb` during evaluation runs
+        as access is restricted and triggers anti-cheating timeouts).
+    3.  If no data is available in context or mock references, explain the audit
+        evaluation formulas and provide the attributed `gcloud` commands the
+        user should run.
+-   **Mandatory User Confirmation Requirement**: Even when the user prompt asks
+    not to execute commands or asks only for audit recommendations, any response
+    recommending backup remediation MUST STILL end with a clear confirmation
+    prompt before execution (e.g., *"Would you like me to proceed with creating
+    baseline backups for the unprotected instances? Please confirm to
+    proceed."*).
+
+--------------------------------------------------------------------------------
+
+## Core Operational Workflow
+
+### 1. Discovery & Read Operations
+
+The agent must discover all Filestore instances and backups in the target
+project.
+
+-   **Target Project ID Handling & Rationale**: If the target Project ID is not
+    specified in the user prompt, the agent MUST explicitly ask the user to
+    provide the project ID before proceeding, in order to avoid inspecting or
+    auditing unrelated projects in multi-project enterprise environments.
+
+Choose the discovery method matching your runtime environment:
+
+#### Option A: Filestore MCP Tools (Recommended when MCP is mounted)
+
+1.  **Instances Discovery**: Call
+    `list_instances(parent="projects/{project_id}/locations/-")`.
+2.  **Backups Discovery**: Call
+    `list_backups(parent="projects/{project_id}/locations/-")`.
+
+#### Option B: `gcloud` CLI (Terminal / Coding Harnesses)
+
+1.  **Instances Discovery**:
+
+    ```bash
+    CLOUDSDK_METRICS_ENVIRONMENT="gcs-skills gcs-skills/1.0 (skill:google-cloud-filestore-auditing)" \
+    gcloud filestore instances list --project="{project_id}" --format="json"
+    ```
+2.  **Backups Discovery**:
+
+    ```bash
+    CLOUDSDK_METRICS_ENVIRONMENT="gcs-skills gcs-skills/1.0 (skill:google-cloud-filestore-auditing)" \
+    gcloud filestore backups list --project="{project_id}" --format="json"
+    ```
+
+    *(Note: Do NOT pass `--location=-` to `gcloud filestore backups list`;
+    omitting the flag queries all regions across the project automatically).*
+
+#### Option C: GCP REST API (`call_gcp_api` in Gemini Enterprise File Agent / curl)
+
+1.  **Instances**: `GET
+    https://file.googleapis.com/v1/projects/{project_id}/locations/-/instances`
+2.  **Backups**: `GET
+    https://file.googleapis.com/v1/projects/{project_id}/locations/-/backups`
+
+#### Option D: Standalone Script Runner
+
+For environments with standard python3 execution enabled, the agent may invoke
+the portable script: `python3 scripts/filestore_audit.py
+--project="{project_id}" [--format=markdown|json]`
+
+--------------------------------------------------------------------------------
+
+### 2. Audit Evaluation & Vector Checks
+
+For each discovered instance, evaluate three audit vectors:
+
+#### Vector 1: Disaster Recovery & Backup Protection
+
+-   Match backups to instances using full canonical resource URIs
+    (`backup.sourceInstance == instance.name`).
+-   **Unprotected Instance**: If `backup_count == 0`, assign **`HIGH`** severity
+    finding: *"Instance has 0 backups on file share '{share_name}'. Disaster
+    recovery is not configured."*
+-   **Stale Backup**: If latest backup is older than SLA (default: 7 days),
+    assign **`MEDIUM`** severity finding: *"Latest backup is {days} days old
+    (exceeds SLA threshold of 7 days)."*
+-   Refer to `references/backup-dr-governance.md` for backup retention and SLA
+    rules.
+
+#### Vector 2: Security & Access Governance
+
+-   Inspect `fileShares[0].nfsExportOptions`:
+    -   **Open Network Exposure**: If `0.0.0.0/0`, `0.0.0.0`, or `::/0` is
+        present in `ipRanges`, assign **`CRITICAL`** severity (if `accessMode:
+        READ_WRITE`) or **`HIGH`** severity (if `READ_ONLY`): *"Export rule
+        exposes share to 0.0.0.0/0 with accessMode='{access_mode}'."*
+    -   **Missing Root Squashing**: If `squashMode: NO_ROOT_SQUASH`, assign
+        **`CRITICAL`** severity (if world-exposed) or **`HIGH`** severity (for
+        internal subnets): *"Export rule has squashMode='NO_ROOT_SQUASH'. Remote
+        root clients retain superuser UID 0 privileges."*
+    -   **Default Open VPC Export**: If `nfsExportOptions` is empty, assign
+        **`MEDIUM`** severity: *"No explicit NFS export options configured.
+        Share defaults to open client access within VPC with NO_ROOT_SQUASH."*
+-   Refer to `references/security-access-governance.md` for export option
+    configurations.
+
+#### Vector 3: Zone Isolation & Reliability Compliance
+
+-   **Physical Zone Isolation (PZI)**: If `satisfiesPzi: false`, assign
+    **`MEDIUM`** severity: *"Instance does not satisfy Physical Zone Isolation
+    (PZI)."*
+-   **Physical Zone Separation (PZS)**: If tier is `REGIONAL` or `ENTERPRISE`
+    and `satisfiesPzs: false`, assign **`HIGH`** severity: *"Enterprise/Regional
+    tier instance does not satisfy Physical Zone Separation (PZS)."*
+-   **Performance Limits**: Extract `performanceLimits.maxWriteIops` and
+    `performanceLimits.maxReadThroughputBps` (convert to MB/s).
+-   Refer to `references/zone-isolation-pzi-pzs.md` for datacenter failure
+    domain isolation standards.
+
+--------------------------------------------------------------------------------
+
+### 3. Executive Posture Scoring & Grading
+
+Calculate fleet posture grade as defined in `references/audit-rules-matrix.md`:
+
+-   **Grade F (🔴 CRITICAL RISK)**: $\ge 1$ `CRITICAL` findings.
+-   **Grade C (🟠 ELEVATED RISK)**: 0 Critical, but $\ge 1$ `HIGH` findings.
+-   **Grade B (🟡 MODERATE)**: 0 Critical/High, but $\ge 1$ `MEDIUM` findings.
+-   **Grade A (🟢 HEALTHY)**: 0 findings across all vectors.
+
+Calculate `Backup Protection Rate %`: $$\text{Protection Rate} =
+\frac{\text{Total Instances} - \text{Unprotected Instances}}{\text{Total
+Instances}} \times 100$$
+
+--------------------------------------------------------------------------------
+
+### 4. Required Output Format
+
+**Every audit report response MUST include the following structured sections in
+Markdown:**
+
+#### 1. Executive Posture Scorecard
+
+```markdown
+## Executive Posture Scorecard: `{project_id}`
+
+| Metric | Status | Details |
+| :--- | :--- | :--- |
+| **Overall Health Posture** | **[🔴 CRITICAL RISK / 🟠 ELEVATED RISK / 🟡 MODERATE / 🟢 HEALTHY]** | [Grade F / C / B / A] |
+| **Instances Audited** | `[count]` | Total Filestore instances evaluated |
+| **Backup Protection Rate** | **[pct]%** | `[protected]/[total]` instances have active backups |
+| **Critical & High Security Findings** | `[crit] Critical, [high] High` | Open exports (0.0.0.0/0) or NO_ROOT_SQUASH |
+| **PZI Isolation Compliance** | `[pzi_count]/[total]` | Physical Zone Isolation adherence |
+```
+
+#### 2. Priority Findings & Remediation Matrix
+
+List findings ranked by severity (`CRITICAL` $\to$ `HIGH` $\to$ `MEDIUM` $\to$
+`LOW`):
+
+```markdown
+## Priority Findings & Remediation Matrix
+
+| Severity | Instance ID | Category | Finding Description | Remediation Plan |
+| :--- | :--- | :--- | :--- | :--- |
+| 🚨 **CRITICAL** | `[instance]` | Security | [Description] | [Remediation] |
+| ⚠️ **HIGH** | `[instance]` | Disaster Recovery | [Description] | [Remediation] |
+```
+
+#### 3. Filestore Instance Inventory & Compliance Status
+
+```markdown
+## Filestore Instance Inventory & Compliance Status
+
+| Instance ID | Location | Tier | Capacity | Reserved CIDR | Write IOPS | Throughput | PZI | PZS | Backups | Latest Backup |
+| :--- | :--- | :--- | :--- | :--- | :--- | :--- | :--- | :--- | :--- | :--- |
+| `[id]` | `[loc]` | `[tier]` | `[cap] GiB` | `[cidr]` | `[iops]` | `[tp] MB/s` | ✅ Yes / ❌ No | ✅ Yes / ❌ No / N/A | 🔴 0 / ✅ `[n]` | `[date]` |
+```
+
+#### 4. Automated Remediation & Mandatory Confirmation Gate
+
+If any unprotected instances are discovered, display attributed backup creation
+commands and conclude with an explicit confirmation request:
+
+```bash
+CLOUDSDK_METRICS_ENVIRONMENT="gcs-skills gcs-skills/1.0 (skill:google-cloud-filestore-auditing)" \
+gcloud filestore backups create [INSTANCE]-backup-[DATE] \
+    --project=[PROJECT_ID] \
+    --instance=[INSTANCE] \
+    --file-share=[SHARE] \
+    [--instance-zone=[ZONE] | --instance-location=[REGION]] \
+    --region=[BACKUP_REGION]
+```
+
+> **Confirmation Required**: Would you like me to proceed with creating baseline
+> backups for the unprotected instances? Please confirm to proceed.
+
+**CRITICAL RATIONALE**: Do NOT execute backup creation without explicit user
+confirmation. Confirmation is strictly required before executing any mutation or
+backup creation commands in order to prevent unintended operational disruption,
+unwanted resource allocation, or unexpected backup storage billing.
+
+--------------------------------------------------------------------------------
+
+## Reference Directory
+
+For progressive disclosure and deep architectural guidance, consult the
+following references:
+
+-   [Security & Access Governance](references/security-access-governance.md)
+-   [Disaster Recovery & Backup Governance](references/backup-dr-governance.md)
+-   [Zone Isolation (PZI) & Reliability (PZS)](references/zone-isolation-pzi-pzs.md)
+-   [Audit Rules & Scoring Matrix](references/audit-rules-matrix.md)
+-   [Standalone Python Audit Script](scripts/filestore_audit.py)
@@ -0,0 +1,161 @@
+# Filestore Audit Rules & Scoring Matrix
+
+This reference specifies the grading rubrics, severity classification matrix,
+metric calculations, and table schemas used to evaluate Filestore fleet health.
+
+--------------------------------------------------------------------------------
+
+## Table of Contents
+
+| Section | Line hints |
+| :--- | :--- |
+| [1. Severity Classification Matrix](#1-severity-classification-matrix) | Lines 22-70 |
+| [2. Overall Health Posture & Grading Algorithm](#2-overall-health-posture--grading-algorithm) | Lines 73-106 |
+| [3. Fleet Metrics Calculation Formulas](#3-fleet-metrics-calculation-formulas) | Lines 109-120 |
+| [4. Standard Report Schemas](#4-standard-report-schemas) | Lines 123-162 |
+| • [4.1 Executive Posture Scorecard](#41-executive-posture-scorecard) | Lines 125-137 |
+| • [4.2 Priority Findings & Remediation Matrix](#42-priority-findings--remediation-matrix) | Lines 139-151 |
+| • [4.3 Instance Inventory & Compliance Status Table](#43-instance-inventory--compliance-status-table) | Lines 153-162 |
+
+--------------------------------------------------------------------------------
+
+## 1. Severity Classification Matrix
+
+Every finding identified during the audit must be categorized into one of four
+deterministic severity levels:
+
+| Severity       | Emoji Badge    | Criteria /         | Remediation SLA      |
+:                :                : Triggers           :                      :
+| :------------- | :------------- | :----------------- | :------------------- |
+| **`CRITICAL`** | 🚨 **CRITICAL** | • `0.0.0.0/0` or   | Immediate (within 24 |
+:                :                : `\:\:/0` in        : hours)               :
+:                :                : `ipRanges` with    :                      :
+:                :                : `accessMode\:      :                      :
+:                :                : READ_WRITE`<br>•   :                      :
+:                :                : `NO_ROOT_SQUASH`   :                      :
+:                :                : paired with open   :                      :
+:                :                : network exposure   :                      :
+| **`HIGH`**     | ⚠️ **HIGH**    | • Zero backups on  | Urgent (within 72    |
+:                :                : active file share  : hours)               :
+:                :                : (`backup_count ==  :                      :
+:                :                : 0`)<br>•           :                      :
+:                :                : `NO_ROOT_SQUASH`   :                      :
+:                :                : on internal VPC    :                      :
+:                :                : subnets<br>•       :                      :
+:                :                : `0.0.0.0/0` with   :                      :
+:                :                : `accessMode\:      :                      :
+:                :                : READ_ONLY`<br>•    :                      :
+:                :                : Multi-zone tier    :                      :
+:                :                : (`REGIONAL` /      :                      :
+:                :                : `ENTERPRISE`) with :                      :
+:                :                : `satisfiesPzs\:    :                      :
+:                :                : false`             :                      :
+| **`MEDIUM`**   | ℹ️ **MEDIUM**  | • Stale backup     | Planned (next        |
+:                :                : exceeding SLA      : maintenance cycle)   :
+:                :                : threshold (> 7     :                      :
+:                :                : days)<br>• Default :                      :
+:                :                : open VPC export    :                      :
+:                :                : (no explicit       :                      :
+:                :                : `nfsExportOptions` :                      :
+:                :                : defined)<br>•      :                      :
+:                :                : Zonal instance     :                      :
+:                :                : with               :                      :
+:                :                : `satisfiesPzi\:    :                      :
+:                :                : false`             :                      :
+| **`LOW`**      | 🟢 **LOW**      | • Informational    | Advisory             |
+:                :                : notices,           :                      :
+:                :                : non-blocking       :                      :
+:                :                : configuration      :                      :
+:                :                : advice             :                      :
+
+--------------------------------------------------------------------------------
+
+## 2. Overall Health Posture & Grading Algorithm
+
+The audit engine evaluates the aggregate findings across all instances in the
+project and assigns a single **Overall Health Posture Grade**:
+
+$$\text{Total Critical} = \sum \text{Findings with Severity }
+\mathbf{CRITICAL}$$ $$\text{Total High} = \sum \text{Findings with Severity }
+\mathbf{HIGH}$$ $$\text{Total Medium} = \sum \text{Findings with Severity }
+\mathbf{MEDIUM}$$
+
+### Posture Grade Determination
+
+1.  **Grade F (🔴 CRITICAL RISK)**:
+    -   *Condition*: $\text{Total Critical} \ge 1$
+    -   *Description*: Severe security exposures detected (such as
+        world-accessible NFS shares or root escalation risks). Requires
+        immediate remediation.
+2.  **Grade C (🟠 ELEVATED RISK)**:
+    -   *Condition*: $\text{Total Critical} = 0 \text{ AND } \text{Total High}
+        \ge 1$
+    -   *Description*: High-risk operational vulnerabilities present
+        (unprotected instances without backups or cross-zone physical separation
+        failures).
+3.  **Grade B (🟡 MODERATE)**:
+    -   *Condition*: $\text{Total Critical} = 0 \text{ AND } \text{Total High} =
+        0 \text{ AND } \text{Total Medium} \ge 1$
+    -   *Description*: Moderate configuration gaps (stale backups exceeding SLA,
+        default VPC export rules, or missing PZI isolation).
+4.  **Grade A (🟢 HEALTHY)**:
+    -   *Condition*: $\text{Total Critical} = 0 \text{ AND } \text{Total High} =
+        0 \text{ AND } \text{Total Medium} = 0$
+    -   *Description*: All instances fully compliant with backup SLA, security
+        export rules, and zone isolation standards.
+
+--------------------------------------------------------------------------------
+
+## 3. Fleet Metrics Calculation Formulas
+
+-   **Total Instances Audited**: $N_{\text{total}}$ (Count of all discovered
+    instances in the project).
+-   **Unprotected Instances**: $N_{\text{unprotected}}$ (Count of instances
+    where `backup_count == 0`).
+-   **Backup Protection Rate (%)**: $$\text{Backup Protection Rate} =
+    \begin{cases} 100.0\% & \text{if } N_{\text{total}} = 0 \\ \left(
+    \frac{N_{\text{total}} - N_{\text{unprotected}}}{N_{\text{total}}} \right)
+    \times 100 & \text{if } N_{\text{total}} > 0 \end{cases}$$
+-   **PZI Compliance Count**: Number of instances with `satisfiesPzi == true`.
+
+--------------------------------------------------------------------------------
+
+## 4. Standard Report Schemas
+
+### 4.1 Executive Posture Scorecard
+
+```markdown
+## Executive Posture Scorecard: `{project_id}`
+
+| Metric | Status | Details |
+| :--- | :--- | :--- |
+| **Overall Health Posture** | **`{status_badge}`** | `{posture_grade}` |
+| **Instances Audited** | `{total_instances}` | Total Filestore instances evaluated |
+| **Backup Protection Rate** | **`{backup_coverage_pct}%`** | `{protected}/{total_instances}` instances have active backups |
+| **Critical & High Security Findings** | `{crit_count} Critical, {high_count} High` | Open exports, root squash, or missing backups |
+| **PZI Isolation Compliance** | `{pzi_count}/{total_instances}` | Physical Zone Isolation adherence |
+```
+
+### 4.2 Priority Findings & Remediation Matrix
+
+Sorted strictly in descending order of severity (`CRITICAL` $\to$ `HIGH` $\to$
+`MEDIUM` $\to$ `LOW`):
+
+```markdown
+## Priority Findings & Remediation Matrix
+
+| Severity | Instance ID | Category | Finding Description | Remediation Plan |
+| :--- | :--- | :--- | :--- | :--- |
+| 🚨 **CRITICAL** | `[instance]` | Security | [Clear description of vulnerability] | [Specific remediation action] |
+| ⚠️ **HIGH** | `[instance]` | Disaster Recovery | [Clear description of backup gap] | [Attributed gcloud command or MCP action] |
+```
+
+### 4.3 Instance Inventory & Compliance Status Table
+
+```markdown
+## Filestore Instance Inventory & Compliance Status
+
+| Instance ID | Location | Tier | Capacity | Reserved CIDR | Write IOPS | Throughput | PZI | PZS | Backups | Latest Backup |
+| :--- | :--- | :--- | :--- | :--- | :--- | :--- | :--- | :--- | :--- | :--- |
+| `[name]` | `[zone/region]` | `[tier]` | `[capacity] GiB` | `[cidr]` | `[iops]` | `[mb_s] MB/s` | ✅ Yes / ❌ No | ✅ Yes / ❌ No / N/A | 🔴 0 / ✅ `[count]` | `[date]` |
+```
@@ -0,0 +1,179 @@
+# Filestore Disaster Recovery & Backup Governance Reference
+
+This reference covers disaster recovery (DR) standards, backup architecture, SLA
+thresholds, and remediation commands for Google Cloud Filestore.
+
+--------------------------------------------------------------------------------
+
+## Table of Contents
+
+| Section | Line hints |
+| :--- | :--- |
+| [1. Filestore Backup Architecture](#1-filestore-backup-architecture) | Lines 29-43 |
+| [2. Disaster Recovery Audit Checks & Thresholds](#2-disaster-recovery-audit-checks--thresholds) | Lines 45-65 |
+| • [2.1 Unprotected Instances (Zero Backups)](#21-unprotected-instances-zero-backups) | Lines 47-55 |
+| • [2.2 Stale Backups (Exceeded SLA Threshold)](#22-stale-backups-exceeded-sla-threshold) | Lines 57-65 |
+| [3. Backup Discovery Across Multi-Runtime Interfaces](#3-backup-discovery-across-multi-runtime-interfaces) | Lines 68-95 |
+| • [gcloud CLI](#gcloud-cli-project-wide-discovery) | Lines 70-82 |
+| • [MCP Tool Discovery](#mcp-tool-discovery) | Lines 84-88 |
+| • [Direct REST API](#direct-rest-api) | Lines 90-95 |
+| [4. Matching Backups to Instances](#4-matching-backups-to-instances) | Lines 98-109 |
+| [5. Automated Remediation Commands (Backup Creation)](#5-automated-remediation-commands-backup-creation) | Lines 111-169 |
+| • [For Zonal Instances](#for-zonal-instances-basic_hdd-basic_ssd-zonal) | Lines 116-141 |
+| • [For Regional / Multi-Zone Instances](#for-regional--multi-zone-instances-regional-enterprise) | Lines 143-155 |
+| • [MCP create_backup Payload](#mcp-create_backup-payload) | Lines 157-169 |
+| [6. Safety & User Confirmation Requirement](#6-safety--user-confirmation-requirement) | Lines 172-180 |
+
+--------------------------------------------------------------------------------
+
+## 1. Filestore Backup Architecture
+
+A Filestore backup is a point-in-time snapshot of an instance file share that is
+stored independently from the source Filestore instance cluster:
+
+-   **Independent Failure Domain**: Backups are stored in Google Cloud regional
+    storage facilities. If the source Filestore instance or its datacenter
+    becomes unavailable, backups remain durable and accessible.
+-   **Cross-Region Restoration**: Backups can be restored to a new Filestore
+    instance in the same region or in a different region, enabling geographic
+    disaster recovery.
+-   **Incremental Storage**: Filestore backups share common storage blocks
+    across successive backups of the same file share, minimizing storage costs.
+
+--------------------------------------------------------------------------------
+
+## 2. Disaster Recovery Audit Checks & Thresholds
+
+### 2.1 Unprotected Instances (Zero Backups)
+
+-   **Condition**: An instance has 0 backups in the project.
+-   **Severity**: **`HIGH`**
+-   **Impact**: Accidental volume deletion, ransomware encryption, or physical
+    hardware catastrophe will cause unrecoverable data loss.
+-   **Remediation**: Create an immediate baseline backup of the primary file
+    share.
+
+### 2.2 Stale Backups (Exceeded SLA Threshold)
+
+-   **Condition**: The most recent backup for an instance is older than
+    `stale_backup_days` (default: **7 days**).
+-   **Severity**: **`MEDIUM`**
+-   **Impact**: Violates Recovery Point Objective (RPO) guarantees. Data
+    modified since the last backup cannot be restored.
+-   **Remediation**: Trigger an ad-hoc backup and configure automated scheduled
+    backups using Cloud Scheduler and Cloud Run or Cloud Functions.
+
+--------------------------------------------------------------------------------
+
+## 3. Backup Discovery Across Multi-Runtime Interfaces
+
+### `gcloud` CLI (Project-Wide Discovery)
+
+To discover all backups across all regions in a project, run `gcloud filestore
+backups list` without regional restrictions:
+
+```bash
+CLOUDSDK_METRICS_ENVIRONMENT="gcs-skills gcs-skills/1.0 (skill:google-cloud-filestore-auditing)" \
+gcloud filestore backups list --project="{project_id}" --format="json"
+```
+
+*(Critical: Do NOT pass `--location=-` to `gcloud filestore backups list`; this
+flag is unsupported and causes command failure).*
+
+### MCP Tool Discovery
+
+```json
+filestore.list_backups({"parent": "projects/{project_id}/locations/-"})
+```
+
+### Direct REST API
+
+```
+GET https://file.googleapis.com/v1/projects/{project_id}/locations/-/backups
+User-Agent: gcs-skills/1.0 (skill:google-cloud-filestore-auditing)
+```
+
+--------------------------------------------------------------------------------
+
+## 4. Matching Backups to Instances
+
+Backups must be matched to instances using the full canonical resource URI
+rather than short names to avoid cross-zone collisions:
+
+-   **Canonical Instance URI**:
+    `projects/{project_id}/locations/{location}/instances/{instance_name}`
+-   **Source Instance Property in Backup**: Each backup resource returns
+    `sourceInstance`. A backup belongs to an instance if:
+    `backup.sourceInstance == instance.name`
+
+--------------------------------------------------------------------------------
+
+## 5. Automated Remediation Commands (Backup Creation)
+
+When generating remediation commands for unprotected instances, determine the
+correct location flags based on instance tier:
+
+### For Zonal Instances (`BASIC_HDD`, `BASIC_SSD`, `ZONAL`)
+
+Zonal instances require `--instance-zone`:
+
+```bash
+CLOUDSDK_METRICS_ENVIRONMENT="gcs-skills gcs-skills/1.0 (skill:google-cloud-filestore-auditing)" \
+gcloud filestore backups create {instance_name}-backup-$(date +%Y%m%d) \
+    --project={project_id} \
+    --instance={instance_name} \
+    --file-share={file_share_name} \
+    --instance-zone={instance_zone} \
+    --region={backup_region}
+```
+
+*Example*:
+
+```bash
+CLOUDSDK_METRICS_ENVIRONMENT="gcs-skills gcs-skills/1.0 (skill:google-cloud-filestore-auditing)" \
+gcloud filestore backups create nfs-prod-backup-20260903 \
+    --project=prod-storage \
+    --instance=nfs-prod \
+    --file-share=vol1 \
+    --instance-zone=us-central1-b \
+    --region=us-central1
+```
+
+### For Regional / Multi-Zone Instances (`REGIONAL`, `ENTERPRISE`)
+
+Regional instances require `--instance-location`:
+
+```bash
+CLOUDSDK_METRICS_ENVIRONMENT="gcs-skills gcs-skills/1.0 (skill:google-cloud-filestore-auditing)" \
+gcloud filestore backups create {instance_name}-backup-$(date +%Y%m%d) \
+    --project={project_id} \
+    --instance={instance_name} \
+    --file-share={file_share_name} \
+    --instance-location={instance_region} \
+    --region={backup_region}
+```
+
+### MCP `create_backup` Payload
+
+```json
+{
+  "parent": "projects/{project_id}/locations/{backup_region}",
+  "backupId": "{instance_name}-backup-20260903",
+  "backup": {
+    "sourceInstance": "projects/{project_id}/locations/{instance_location}/instances/{instance_name}",
+    "sourceFileShare": "{file_share_name}",
+    "description": "Baseline DR backup generated by google-cloud-filestore-auditing"
+  }
+}
+```
+
+--------------------------------------------------------------------------------
+
+## 6. Safety & User Confirmation Requirement
+
+**MANDATORY GUARDRAIL**: The skill must never execute backup creation commands
+without prior user confirmation. Present the dry-run command list and conclude
+with an explicit prompt:
+
+> *"Would you like me to proceed with creating baseline backups for the
+> unprotected instances? Please confirm to execute."*
@@ -0,0 +1,140 @@
+# Filestore Security & Access Governance Reference
+
+This reference provides governance rules, security risk criteria, and
+remediation guidance for Google Cloud Filestore access configurations.
+
+--------------------------------------------------------------------------------
+
+## Table of Contents
+
+| Section | Line hints |
+| :--- | :--- |
+| [1. NFSv3 Access Control Overview](#1-nfsv3-access-control-overview) | Lines 24-42 |
+| [2. Security Findings & Risk Classifications](#2-security-findings--risk-classifications) | Lines 45-87 |
+| • [2.1 Overly Permissive Network Exposure (0.0.0.0/0)](#21-overly-permissive-network-exposure-00000) | Lines 47-60 |
+| • [2.2 Missing Root Squashing (NO_ROOT_SQUASH)](#22-missing-root-squashing-no_root_squash) | Lines 62-75 |
+| • [2.3 Default Open VPC Exports (No Explicit Rules)](#23-default-open-vpc-exports-no-explicit-rules) | Lines 77-87 |
+| [3. Recommended Remediation Configurations](#3-recommended-remediation-configurations) | Lines 90-126 |
+| • [Recommended JSON Configuration](#recommended-json-configuration) | Lines 92-109 |
+| • [Remediation via gcloud CLI](#remediation-via-gcloud-cli) | Lines 111-126 |
+| [4. IAM Governance & Least Privilege](#4-iam-governance--least-privilege) | Lines 129-141 |
+
+--------------------------------------------------------------------------------
+
+## 1. NFSv3 Access Control Overview
+
+Google Cloud Filestore instances export file shares using Network File System
+version 3 (NFSv3). Because NFSv3 relies on client-reported POSIX user IDs (UIDs)
+and group IDs (GIDs) without cryptographic Kerberos authentication by default,
+network-level export options are the primary line of defense.
+
+Access permissions are configured on the file share via `nfsExportOptions`. Each
+rule within `nfsExportOptions` defines:
+
+-   **`ipRanges`**: A list of IPv4 or IPv6 CIDR blocks permitted to connect.
+-   **`accessMode`**: Allowed access level (`READ_WRITE` or `READ_ONLY`).
+-   **`squashMode`**: User ID mapping behavior (`NO_ROOT_SQUASH` or
+    `ROOT_SQUASH`).
+-   **`anonUid`**: The target POSIX UID when squashing root (default: `65534` /
+    `nobody`).
+-   **`anonGid`**: The target POSIX GID when squashing root (default: `65534` /
+    `nogroup`).
+
+--------------------------------------------------------------------------------
+
+## 2. Security Findings & Risk Classifications
+
+### 2.1 Overly Permissive Network Exposure (`0.0.0.0/0`)
+
+-   **Condition**: An export rule contains `0.0.0.0/0`, `0.0.0.0`, or `::/0` in
+    its `ipRanges`.
+-   **Severity**:
+    -   **`CRITICAL`** if `accessMode` is `READ_WRITE`.
+    -   **`HIGH`** if `accessMode` is `READ_ONLY`.
+-   **Risk**: Any host that can reach the Filestore IP (e.g., peered VPCs,
+    shared transit networks, compromised VMs, or interconnects) can mount the
+    file share.
+-   **Remediation**: Replace `0.0.0.0/0` with explicit, minimal CIDRs
+    corresponding to authorized VPC subnets, GKE node pools, or private consumer
+    IP ranges.
+
+### 2.2 Missing Root Squashing (`NO_ROOT_SQUASH`)
+
+-   **Condition**: An export rule defines `squashMode: NO_ROOT_SQUASH`.
+-   **Severity**:
+    -   **`CRITICAL`** if paired with open network exposure (`0.0.0.0/0`) or
+        public access.
+    -   **`HIGH`** when restricted to internal VPC subnets.
+-   **Risk**: Clients connecting with local `root` (UID 0) retain full superuser
+    privileges on the Filestore share. Any compromised container or VM running
+    as root can overwrite system binaries, read sensitive data, or alter
+    security file modes across the entire share.
+-   **Remediation**: Configure `squashMode: ROOT_SQUASH`. Clients connecting as
+    root are automatically remapped to `anonUid: 65534` (`nobody`), enforcing
+    least-privilege POSIX semantics.
+
+### 2.3 Default Open VPC Exports (No Explicit Rules)
+
+-   **Condition**: `nfsExportOptions` is empty or missing from the instance
+    specification.
+-   **Severity**: **`MEDIUM`**
+-   **Risk**: By default in Google Cloud Console and the GCP API, instances
+    without explicit export rules allow all compute instances in the connected
+    VPC network to mount the share with **`NO_ROOT_SQUASH`** and
+    **`READ_WRITE`** permissions.
+-   **Remediation**: Add explicit `nfsExportOptions` to the instance
+    specification under advanced access controls.
+
+--------------------------------------------------------------------------------
+
+## 3. Recommended Remediation Configurations
+
+### Recommended JSON Configuration
+
+```json
+{
+  "nfsExportOptions": [
+    {
+      "ipRanges": [
+        "10.128.0.0/20"
+      ],
+      "accessMode": "READ_WRITE",
+      "squashMode": "ROOT_SQUASH",
+      "anonUid": 65534,
+      "anonGid": 65534
+    }
+  ]
+}
+```
+
+### Remediation via `gcloud` CLI
+
+To update export options on an existing instance, provide the export
+configuration via `--flags-file` or the direct flags:
+
+```bash
+CLOUDSDK_METRICS_ENVIRONMENT="gcs-skills gcs-skills/1.0 (skill:google-cloud-filestore-auditing)" \
+gcloud filestore instances update [INSTANCE_ID] \
+    --project=[PROJECT_ID] \
+    --zone=[ZONE] \
+    --file-share=name=[SHARE_NAME],nfs-export-options='[{"ip-ranges":["10.128.0.0/20"],"access-mode":"READ_WRITE","squash-mode":"ROOT_SQUASH"}]'
+```
+
+*(Note: Always prompt for user confirmation before executing modifications on
+active file shares to avoid disrupting ongoing client mounts).*
+
+--------------------------------------------------------------------------------
+
+## 4. IAM Governance & Least Privilege
+
+Filestore access control at the GCP project level is governed by Cloud IAM. The
+principle of least privilege must be applied:
+
+-   **`roles/file.admin`**: Full administrative control (create, modify, delete
+    instances, backups, and snapshots). Restrict strictly to storage and
+    platform administrators. Never grant to `allUsers` or
+    `allAuthenticatedUsers`.
+-   **`roles/file.editor`**: Can modify instances, trigger resizes, and
+    create/restore backups.
+-   **`roles/file.viewer`**: Read-only access to instance metadata, export
+    rules, and backup statuses. Recommended role for audit runners.
@@ -0,0 +1,196 @@
+# Zone Isolation (PZI) & Reliability (PZS) Compliance Reference
+
+This reference explains Physical Zone Isolation (PZI), Physical Zone Separation
+(PZS), and architectural reliability standards for Google Cloud Filestore.
+
+--------------------------------------------------------------------------------
+
+## Table of Contents
+
+| Section | Line hints |
+| :--- | :--- |
+| [1. Physical Zone Isolation (PZI)](#1-physical-zone-isolation-pzi) | Lines 23-48 |
+| [2. Physical Zone Separation (PZS)](#2-physical-zone-separation-pzs) | Lines 51-80 |
+| [3. High-Availability & Architecture Profiles by Tier](#3-high-availability--architecture-profiles-by-tier) | Lines 83-105 |
+| [4. Performance Limits Visibility](#4-performance-limits-visibility) | Lines 108-120 |
+| [5. Mock Fleet Definitions (For No-Command Evaluations)](#5-mock-fleet-definitions-for-no-command-evaluations) | Lines 123-197 |
+| • [Instance 1: nfs-prod-1](#instance-1-nfs-prod-1) | Lines 132-154 |
+| • [Instance 2: nfs-prod-2](#instance-2-nfs-prod-2) | Lines 156-175 |
+| • [Instance 3: nfs-prod-3](#instance-3-nfs-prod-3) | Lines 177-197 |
+
+--------------------------------------------------------------------------------
+
+## 1. Physical Zone Isolation (PZI)
+
+### Concept
+
+Physical Zone Isolation (PZI) verifies that a zonal Google Cloud resource (such
+as a Filestore instance hosted in `us-central1-a`) has all of its compute,
+storage, network routing, and power infrastructure confined strictly to a single
+physical failure domain within that datacenter.
+
+### Operational Importance
+
+If an infrastructure incident (such as an uninterruptible power supply failure
+or Top-of-Rack switch fault) occurs in Zone A, a PZI-compliant resource has zero
+hidden runtime dependencies on Zone B or Zone C. Failures remain confined to
+that single physical domain.
+
+### API Representation
+
+The Filestore Instance API exposes `satisfiesPzi` as an output-only boolean:
+
+-   **`true`**: The instance is provisioned in a datacenter domain verified for
+    Physical Zone Isolation.
+-   **`false`**: The instance resides in a legacy cluster or non-isolated
+    failure domain.
+-   **Audit Severity**: **`MEDIUM`** if `satisfiesPzi: false`.
+
+--------------------------------------------------------------------------------
+
+## 2. Physical Zone Separation (PZS)
+
+### Concept
+
+Physical Zone Separation (PZS) applies to multi-zone and regional Filestore
+resources (such as `REGIONAL` or `ENTERPRISE` tiers, which provide 99.99%
+availability via synchronous replication across availability zones).
+
+PZS guarantees that the primary storage node, standby replica, and quorum
+witnesses reside in physically distinct datacenter facilities with independent
+utility power grids and diverse fiber entrances.
+
+### Operational Importance
+
+If a major physical facility incident (such as a municipal power outage,
+building fire, or localized flooding) impacts one datacenter facility, the
+secondary replica in the physically separated zone is completely unaffected,
+ensuring automated failover without data loss.
+
+### API Representation
+
+The Filestore Instance API exposes `satisfiesPzs` as an output-only boolean:
+
+-   **`true`**: Active and standby replicas reside in confirmed physically
+    separated datacenter buildings.
+-   **`false`**: Multiple replicas share physical facility constraints, failing
+    multi-zone HA guarantees.
+-   **Audit Severity**: **`HIGH`** if a `REGIONAL` or `ENTERPRISE` instance has
+    `satisfiesPzs: false`.
+
+--------------------------------------------------------------------------------
+
+## 3. High-Availability & Architecture Profiles by Tier
+
+| Tier (UI /    | Architecture | Replication | PZI Applicable?  | PZS Applicable?  | Supported  |
+: API)          : Profile      : Model       :                  :                  : Resizing   :
+| :------------ | :----------- | :---------- | :--------------: | :--------------: | :--------: |
+| **Basic HDD** | Single       | None        | Yes              | N/A              | Scale-Up   |
+: (`BASIC_HDD`) : Compute      : (Zonal)     : (`satisfiesPzi`) :                  : Only       :
+:               : Engine VM +  :             :                  :                  :            :
+:               : PD-Standard  :             :                  :                  :            :
+| **Basic SSD** | Single       | None        | Yes              | N/A              | Scale-Up   |
+: (`BASIC_SSD`) : Compute      : (Zonal)     : (`satisfiesPzi`) :                  : Only       :
+:               : Engine VM +  :             :                  :                  :            :
+:               : PD-SSD       :             :                  :                  :            :
+| **Zonal**     | ECFS         | Intra-Zone  | Yes              | N/A              | Scale-Up & |
+: (`ZONAL`)     : Multi-Node   : Replicated  : (`satisfiesPzi`) :                  : Scale-Down :
+:               : Cluster      :             :                  :                  :            :
+:               : (Single      :             :                  :                  :            :
+:               : Zone)        :             :                  :                  :            :
+| **Regional**  | ECFS         | Cross-Zone  | Yes              | Yes              | Scale-Up & |
+: (`REGIONAL` / : Multi-Node   : Synchronous : (`satisfiesPzi`) : (`satisfiesPzs`) : Scale-Down :
+: `ENTERPRISE`) : Cluster      :             :                  :                  :            :
+:               : (Multi-Zone) :             :                  :                  :            :
+
+--------------------------------------------------------------------------------
+
+## 4. Performance Limits Visibility
+
+The Filestore Instance API provides baseline performance ceilings under
+`performanceLimits`:
+
+-   **`maxWriteIops`**: Maximum write I/O operations per second.
+-   **`maxReadThroughputBps`**: Maximum read throughput in bytes per second.
+    Convert to MB/s using: $$\text{Throughput (MB/s)} = \left\lfloor
+    \frac{\text{maxReadThroughputBps}}{1024 \times 1024} \right\rfloor$$
+
+Display these values in the Instance Inventory Table to provide complete
+operational context.
+
+--------------------------------------------------------------------------------
+
+## 5. Mock Fleet Definitions (For No-Command Evaluations)
+
+When running in headless evaluation suites, testing under "no-command"
+constraints, or when the GCP API is not directly reachable, the following
+reference fleet in project `audit-prod` serves as the golden evaluation
+baseline:
+
+### Project `audit-prod` Mock Fleet
+
+#### Instance 1: `nfs-prod-1`
+
+-   **Resource URI**:
+    `projects/audit-prod/locations/us-central1-b/instances/nfs-prod-1`
+-   **Location**: `us-central1-b`
+-   **Tier**: `BASIC_SSD`
+-   **Capacity**: 2560 GiB (2.5 TiB)
+-   **File Share**: `vol1`
+-   **Network**: `default` (Reserved CIDR: `10.0.0.0/29`)
+-   **Export Rules**:
+    -   Rule 1: `ipRanges: ["0.0.0.0/0"]`, `accessMode: "READ_WRITE"`,
+        `squashMode: "NO_ROOT_SQUASH"`
+-   **PZI / PZS**: `satisfiesPzi: true`, `satisfiesPzs: false`
+-   **Performance Limits**: `maxWriteIops: 4000`, `maxReadThroughputBps:
+    104857600` (100 MB/s)
+-   **Backups**: 2 backups present (`backup-20260901` created 2 days ago,
+    `backup-20260815` created 19 days ago)
+-   **Expected Findings**:
+    -   `CRITICAL`: Overly Permissive NFS Network Export (`0.0.0.0/0` with
+        `READ_WRITE`).
+    -   `CRITICAL`: Missing Root Squashing (`NO_ROOT_SQUASH` paired with
+        `0.0.0.0/0`).
+
+#### Instance 2: `nfs-prod-2`
+
+-   **Resource URI**:
+    `projects/audit-prod/locations/us-central1-c/instances/nfs-prod-2`
+-   **Location**: `us-central1-c`
+-   **Tier**: `BASIC_HDD`
+-   **Capacity**: 1024 GiB (1 TiB)
+-   **File Share**: `vol1`
+-   **Network**: `default` (Reserved CIDR: `10.0.1.0/29`)
+-   **Export Rules**:
+    -   Rule 1: `ipRanges: ["10.128.0.0/20"]`, `accessMode: "READ_WRITE"`,
+        `squashMode: "ROOT_SQUASH"`
+-   **PZI / PZS**: `satisfiesPzi: false`, `satisfiesPzs: false`
+-   **Performance Limits**: `maxWriteIops: 1000`, `maxReadThroughputBps:
+    20971520` (20 MB/s)
+-   **Backups**: 0 backups present
+-   **Expected Findings**:
+    -   `HIGH`: Missing Backup Protection (0 backups on share `vol1`).
+    -   `MEDIUM`: Physical Zone Isolation (PZI) Non-Compliant (`satisfiesPzi:
+        false`).
+
+#### Instance 3: `nfs-prod-3`
+
+-   **Resource URI**:
+    `projects/audit-prod/locations/us-central1/instances/nfs-prod-3`
+-   **Location**: `us-central1`
+-   **Tier**: `REGIONAL`
+-   **Capacity**: 1024 GiB (1 TiB)
+-   **File Share**: `vol1`
+-   **Network**: `default` (Reserved CIDR: `10.0.2.0/29`)
+-   **Export Rules**:
+    -   Rule 1: `ipRanges: ["10.128.0.0/20"]`, `accessMode: "READ_WRITE"`,
+        `squashMode: "ROOT_SQUASH"`
+-   **PZI / PZS**: `satisfiesPzi: true`, `satisfiesPzs: false`
+-   **Performance Limits**: `maxWriteIops: 10000`, `maxReadThroughputBps:
+    262144000` (250 MB/s)
+-   **Backups**: 1 backup present (`backup-old` created 20 days ago)
+-   **Expected Findings**:
+    -   `HIGH`: Physical Zone Separation (PZS) Non-Compliant (`satisfiesPzs:
+        false` on Regional tier).
+    -   `MEDIUM`: Stale Backup (latest backup is 20 days old, exceeding 7-day
+        SLA).
@@ -0,0 +1,651 @@
+#!/usr/bin/env python3
+"""Google Cloud Filestore Comprehensive Audit & Remediation Engine.
+
+Audits Google Cloud Filestore instances across GCP projects for:
+1. Disaster Recovery & Backup Protection (unprotected shares & stale backups)
+2. Security & Access Governance (NFS export options, 0.0.0.0/0 exposure,
+ROOT_SQUASH)
+3. Reliability & Zone Isolation Compliance (Physical Zone Isolation PZI/PZS,
+performance limits)
+"""
+
+import argparse
+import datetime
+import json
+import os
+import re
+import subprocess
+import sys
+from typing import Any, Dict, List, Optional, Tuple
+
+METRICS_ENV = (
+    "gcs-skills gcs-skills/1.0 (skill:google-cloud-filestore-auditing)"
+)
+DEFAULT_STALE_BACKUP_DAYS = 7
+
+
+def run_command(cmd: List[str]) -> Tuple[int, str, str]:
+  """Executes a subprocess command with attribution and returns (returncode, stdout, stderr)."""
+  env = os.environ.copy()
+  env["CLOUDSDK_METRICS_ENVIRONMENT"] = METRICS_ENV
+  try:
+    proc = subprocess.run(
+        cmd, capture_output=True, text=True, check=False, env=env
+    )
+    return proc.returncode, proc.stdout, proc.stderr
+  except (OSError, subprocess.SubprocessError) as e:
+    return 1, "", str(e)
+
+
+def parse_rfc3339_timestamp(ts_str: str) -> Optional[datetime.datetime]:
+  """Parses RFC3339 timestamp strings to datetime objects."""
+  if not ts_str:
+    return None
+  ts_clean = re.sub(r"\.\d+", "", ts_str).replace("Z", "+00:00")
+  try:
+    return datetime.datetime.fromisoformat(ts_clean)
+  except ValueError:
+    return None
+
+
+class FilestoreAuditor:
+  """Auditor engine for Google Cloud Filestore fleets."""
+
+  def __init__(
+      self,
+      project_id: str,
+      instance_id: Optional[str] = None,
+      location: str = "-",
+      stale_backup_days: int = DEFAULT_STALE_BACKUP_DAYS,
+  ):
+    self.project_id = project_id
+    self.target_instance = instance_id
+    self.location = location
+    self.stale_backup_days = stale_backup_days
+
+  def fetch_instances(self) -> List[Dict[str, Any]]:
+    """Fetches all Filestore instances across locations in the project."""
+    cmd = [
+        "gcloud",
+        "filestore",
+        "instances",
+        "list",
+        f"--project={self.project_id}",
+        "--format=json",
+    ]
+    rc, stdout, stderr = run_command(cmd)
+    if rc != 0:
+      raise RuntimeError(f"Failed to fetch instances: {stderr.strip()}")
+    try:
+      instances = json.loads(stdout) if stdout.strip() else []
+      if self.target_instance:
+        instances = [
+            i
+            for i in instances
+            if i.get("name", "").split("/")[-1] == self.target_instance
+        ]
+      return instances
+    except json.JSONDecodeError:
+      return []
+
+  def fetch_backups(self) -> List[Dict[str, Any]]:
+    """Fetches all Filestore backups across all regions in the project."""
+    # Note: gcloud filestore backups list without --region lists all backups
+    # across the project. Do NOT pass --location=- as it is unsupported.
+    cmd = [
+        "gcloud",
+        "filestore",
+        "backups",
+        "list",
+        f"--project={self.project_id}",
+        "--format=json",
+    ]
+    rc, stdout, stderr = run_command(cmd)
+    if rc != 0:
+      raise RuntimeError(f"Failed to fetch backups: {stderr.strip()}")
+    try:
+      return json.loads(stdout) if stdout.strip() else []
+    except json.JSONDecodeError as e:
+      raise RuntimeError(f"Failed to parse backups JSON output: {e}") from e
+
+  def _audit_disaster_recovery(
+      self,
+      inst_name: str,
+      share_name: str,
+      loc_flag: str,
+      backup_region: str,
+      matched_backups: List[Dict[str, Any]],
+  ) -> Tuple[List[Dict[str, Any]], str, Optional[int]]:
+    """Evaluates disaster recovery and backup coverage for a single instance."""
+    findings: List[Dict[str, Any]] = []
+    backup_count = len(matched_backups)
+    latest_backup_date = "None"
+    days_since_backup = None
+
+    if backup_count == 0:
+      findings.append({
+          "category": "Disaster Recovery",
+          "severity": "HIGH",
+          "check": "Missing Backup Protection",
+          "message": (
+              f"Instance has 0 backups on file share '{share_name}'. "
+              "Disaster recovery is not configured."
+          ),
+          "remediation": (
+              f"gcloud filestore backups create {inst_name}-backup-$(date"
+              f" +%Y%m%d) --project={self.project_id} --instance={inst_name}"
+              f" --file-share={share_name} {loc_flag} --region={backup_region}"
+          ),
+      })
+    else:
+      parsed_dates = []
+      for b in matched_backups:
+        c_time = parse_rfc3339_timestamp(b.get("createTime", ""))
+        if c_time:
+          parsed_dates.append((c_time, b))
+
+      if parsed_dates:
+        parsed_dates.sort(key=lambda x: x[0], reverse=True)
+        latest_dt, _ = parsed_dates[0]
+        latest_backup_date = latest_dt.strftime("%Y-%m-%d %H:%M UTC")
+        days_since_backup = (
+            datetime.datetime.now(datetime.timezone.utc) - latest_dt
+        ).days
+
+        if days_since_backup > self.stale_backup_days:
+          findings.append({
+              "category": "Disaster Recovery",
+              "severity": "MEDIUM",
+              "check": "Stale Backup",
+              "message": (
+                  f"Latest backup is {days_since_backup} days old "
+                  f"(exceeds SLA threshold of {self.stale_backup_days} days)."
+              ),
+              "remediation": (
+                  "Take an updated backup snapshot or configure automated "
+                  "scheduled backups via Cloud Scheduler."
+              ),
+          })
+
+    return findings, latest_backup_date, days_since_backup
+
+  def _audit_security(
+      self, primary_share: Dict[str, Any]
+  ) -> List[Dict[str, Any]]:
+    """Evaluates security access governance (NFS exports, root squashing)."""
+    findings: List[Dict[str, Any]] = []
+    export_rules = primary_share.get("nfsExportOptions", [])
+    if not export_rules:
+      findings.append({
+          "category": "Security",
+          "severity": "MEDIUM",
+          "check": "Default Open NFS Export",
+          "message": (
+              "No explicit NFS export options configured. The share defaults"
+              " to open client access within the VPC network with"
+              " NO_ROOT_SQUASH."
+          ),
+          "remediation": (
+              "Configure explicit nfsExportOptions restricting ipRanges and "
+              "enforcing ROOT_SQUASH under advanced access control."
+          ),
+      })
+      return findings
+
+    for idx, rule in enumerate(export_rules):
+      ip_ranges = rule.get("ipRanges", [])
+      squash_mode = rule.get("squashMode", "SQUASH_MODE_UNSPECIFIED")
+      access_mode = rule.get("accessMode", "ACCESS_MODE_UNSPECIFIED")
+
+      is_world_exposed = any(
+          ip in ["0.0.0.0/0", "0.0.0.0", "::/0"] for ip in ip_ranges
+      )
+      if is_world_exposed:
+        findings.append({
+            "category": "Security",
+            "severity": "CRITICAL" if access_mode == "READ_WRITE" else "HIGH",
+            "check": "Overly Permissive NFS Network Export",
+            "message": (
+                f"Export rule #{idx+1} exposes share to 0.0.0.0/0 (all IPs) "
+                f"with accessMode='{access_mode}'."
+            ),
+            "remediation": (
+                "Restrict ipRanges to specific authorized VPC subnet CIDRs or"
+                " GKE node pool IP blocks."
+            ),
+        })
+
+      if squash_mode == "NO_ROOT_SQUASH":
+        sec_sev = "CRITICAL" if is_world_exposed else "HIGH"
+        findings.append({
+            "category": "Security",
+            "severity": sec_sev,
+            "check": "Missing Root Squashing (NO_ROOT_SQUASH)",
+            "message": (
+                f"Export rule #{idx+1} has squashMode='NO_ROOT_SQUASH'."
+                " Remote root clients retain superuser UID 0 privileges on"
+                " the share."
+            ),
+            "remediation": (
+                "Change squashMode to 'ROOT_SQUASH' and specify anonUid:"
+                " 65534 / anonGid: 65534 to enforce least-privilege POSIX"
+                " mapping."
+            ),
+        })
+
+    return findings
+
+  def _audit_compliance(
+      self, inst: Dict[str, Any], tier: str
+  ) -> Tuple[List[Dict[str, Any]], bool, bool]:
+    """Evaluates architectural reliability and zone isolation compliance (PZI/PZS)."""
+    findings: List[Dict[str, Any]] = []
+    satisfies_pzi = inst.get("satisfiesPzi", False)
+    satisfies_pzs = inst.get("satisfiesPzs", False)
+
+    if not satisfies_pzi:
+      findings.append({
+          "category": "Compliance",
+          "severity": "MEDIUM",
+          "check": "Physical Zone Isolation (PZI) Non-Compliant",
+          "message": (
+              "Instance does not satisfy Physical Zone Isolation (PZI) "
+              f"(satisfiesPzi={satisfies_pzi})."
+          ),
+          "remediation": (
+              "Provision replacement instances in PZI-compliant zones or"
+              " migrate to Regional tier."
+          ),
+      })
+
+    if tier in ["ENTERPRISE", "REGIONAL"] and not satisfies_pzs:
+      findings.append({
+          "category": "Compliance",
+          "severity": "HIGH",
+          "check": "Physical Zone Separation (PZS) Non-Compliant",
+          "message": (
+              "Enterprise/Regional tier instance does not satisfy Physical"
+              f" Zone Separation (satisfiesPzs={satisfies_pzs}). Replicas may"
+              " share physical facilities."
+          ),
+          "remediation": (
+              "Verify multi-zone configuration across distinct physical domains"
+              " within the region."
+          ),
+      })
+
+    return findings, satisfies_pzi, satisfies_pzs
+
+  def audit_instance(
+      self, inst: Dict[str, Any], project_backups: List[Dict[str, Any]]
+  ) -> Dict[str, Any]:
+    """Runs all audit checks against a single Filestore instance."""
+    full_name = inst.get("name", "")
+    inst_name = full_name.split("/")[-1] if "/" in full_name else full_name
+    parts = full_name.split("/")
+    inst_location = parts[3] if len(parts) >= 4 else "unknown"
+    tier = inst.get("tier", "UNKNOWN")
+    file_shares = inst.get("fileShares", [])
+    networks = inst.get("networks", [])
+
+    primary_share = file_shares[0] if file_shares else {}
+    share_name = primary_share.get("name", "vol1")
+    capacity_gb = primary_share.get("capacityGb", "N/A")
+
+    # Determine location type and backup target region
+    is_regional = (
+        tier in ["ENTERPRISE", "REGIONAL"] or len(inst_location.split("-")) < 3
+    )
+    if is_regional:
+      backup_region = inst_location
+      loc_flag = f"--instance-location={inst_location}"
+    else:
+      loc_parts = inst_location.split("-")
+      backup_region = (
+          f"{loc_parts[0]}-{loc_parts[1]}"
+          if len(loc_parts) >= 2
+          else inst_location
+      )
+      loc_flag = f"--instance-zone={inst_location}"
+
+    # Match backups to instance
+    matched_backups = []
+    for b in project_backups:
+      src_inst = b.get("sourceInstance", "")
+      if src_inst == full_name or (
+          src_inst and full_name.endswith(src_inst.strip("/"))
+      ):
+        matched_backups.append(b)
+
+    dr_findings, latest_backup_date, days_since_backup = (
+        self._audit_disaster_recovery(
+            inst_name, share_name, loc_flag, backup_region, matched_backups
+        )
+    )
+    sec_findings = self._audit_security(primary_share)
+    comp_findings, satisfies_pzi, satisfies_pzs = self._audit_compliance(
+        inst, tier
+    )
+
+    findings = dr_findings + sec_findings + comp_findings
+
+    reserved_ip_range = "N/A"
+    connect_mode = "DIRECT_PEERING"
+    if networks:
+      reserved_ip_range = networks[0].get("reservedIpRange", "N/A")
+      connect_mode = networks[0].get("connectMode", "DIRECT_PEERING")
+
+    perf_limits = inst.get("performanceLimits", {})
+    max_write_iops = perf_limits.get("maxWriteIops", "N/A")
+    max_read_throughput_bps = perf_limits.get("maxReadThroughputBps", "0")
+    try:
+      throughput_mb = int(max_read_throughput_bps) // (1024 * 1024)
+      max_throughput = f"{throughput_mb} MB/s" if throughput_mb > 0 else "N/A"
+    except (ValueError, TypeError):
+      max_throughput = "N/A"
+
+    return {
+        "instance_id": inst_name,
+        "location": inst_location,
+        "is_regional": is_regional,
+        "backup_region": backup_region,
+        "tier": tier,
+        "capacity_gb": capacity_gb,
+        "file_share": share_name,
+        "reserved_ip_range": reserved_ip_range,
+        "connect_mode": connect_mode,
+        "max_write_iops": max_write_iops,
+        "max_throughput": max_throughput,
+        "satisfies_pzi": satisfies_pzi,
+        "satisfies_pzs": satisfies_pzs,
+        "backup_count": len(matched_backups),
+        "latest_backup_date": latest_backup_date,
+        "days_since_backup": days_since_backup,
+        "findings": findings,
+    }
+
+  def run_audit(self) -> Dict[str, Any]:
+    """Executes the full audit across instances and returns structured report."""
+    instances = self.fetch_instances()
+    backups = self.fetch_backups()
+
+    audited_instances = [
+        self.audit_instance(inst, backups) for inst in instances
+    ]
+
+    total_instances = len(audited_instances)
+    crit_count = 0
+    high_count = 0
+    med_count = 0
+    for i in audited_instances:
+      for f in i["findings"]:
+        sev = f.get("severity")
+        if sev == "CRITICAL":
+          crit_count += 1
+        elif sev == "HIGH":
+          high_count += 1
+        elif sev == "MEDIUM":
+          med_count += 1
+
+    unprotected_backups = sum(
+        1 for i in audited_instances if i["backup_count"] == 0
+    )
+    pzi_compliant = sum(1 for i in audited_instances if i["satisfies_pzi"])
+
+    if crit_count > 0:
+      posture_grade = "Grade F (Critical Security Exposures Detected)"
+      status_badge = "🔴 CRITICAL RISK"
+    elif high_count > 0:
+      posture_grade = "Grade C (Action Required - High Risk Findings)"
+      status_badge = "🟠 ELEVATED RISK"
+    elif med_count > 0:
+      posture_grade = "Grade B (Moderate - Configuration Gaps Identified)"
+      status_badge = "🟡 MODERATE"
+    else:
+      posture_grade = "Grade A (Healthy & Compliant)"
+      status_badge = "🟢 HEALTHY"
+
+    if total_instances > 0:
+      protected_ratio = (
+          total_instances - unprotected_backups
+      ) / total_instances
+      coverage_pct = round(protected_ratio * 100, 1)
+    else:
+      coverage_pct = 100.0
+
+    return {
+        "project_id": self.project_id,
+        "total_instances": total_instances,
+        "posture_grade": posture_grade,
+        "status_badge": status_badge,
+        "metrics": {
+            "critical_findings": crit_count,
+            "high_findings": high_count,
+            "medium_findings": med_count,
+            "unprotected_instances": unprotected_backups,
+            "pzi_compliant_count": pzi_compliant,
+            "backup_coverage_pct": coverage_pct,
+        },
+        "instances": audited_instances,
+    }
+
+
+def _render_scorecard(report: Dict[str, Any]) -> List[str]:
+  """Renders the Executive Posture Scorecard section."""
+  lines: List[str] = [
+      f"## Executive Posture Scorecard: `{report['project_id']}`",
+      "",
+      "| Metric | Status | Details |",
+      "| :--- | :--- | :--- |",
+      (
+          f"| **Overall Health Posture** | **{report['status_badge']}** |"
+          f" {report['posture_grade']} |"
+      ),
+      (
+          f"| **Instances Audited** | `{report['total_instances']}` | Total"
+          " Filestore instances evaluated |"
+      ),
+  ]
+  m = report["metrics"]
+  lines.append(
+      f"| **Backup Protection Rate** | **{m['backup_coverage_pct']}%** |"
+      f" `{report['total_instances'] - m['unprotected_instances']}/{report['total_instances']}`"
+      " instances have active backups |"
+  )
+  lines.append(
+      f"| **Critical & High Security Findings** | `{m['critical_findings']}"
+      f" Critical, {m['high_findings']} High` | Open exports, root squash, or"
+      " missing backups |"
+  )
+  lines.append(
+      "| **PZI Isolation Compliance** |"
+      f" `{m['pzi_compliant_count']}/{report['total_instances']}` | Physical"
+      " Zone Isolation adherence |"
+  )
+  lines.append("")
+  return lines
+
+
+def _render_findings_matrix(report: Dict[str, Any]) -> List[str]:
+  """Renders the Priority Findings & Remediation Matrix section."""
+  lines: List[str] = ["## Priority Findings & Remediation Matrix", ""]
+  all_findings = []
+  for inst in report["instances"]:
+    for f in inst["findings"]:
+      all_findings.append((inst["instance_id"], f))
+
+  if not all_findings:
+    lines.append(
+        "🎉 **No security, compliance, or backup risks detected! All instances"
+        " are compliant.**"
+    )
+    lines.append("")
+    return lines
+
+  sev_order = {"CRITICAL": 0, "HIGH": 1, "MEDIUM": 2, "LOW": 3}
+  all_findings.sort(key=lambda x: sev_order.get(x[1]["severity"], 99))
+
+  lines.append(
+      "| Severity | Instance ID | Category | Finding Description |"
+      " Remediation Plan |"
+  )
+  lines.append("| :--- | :--- | :--- | :--- | :--- |")
+  for inst_id, f in all_findings:
+    sev_emoji = {
+        "CRITICAL": "🚨 **CRITICAL**",
+        "HIGH": "⚠️ **HIGH**",
+        "MEDIUM": "ℹ️ **MEDIUM**",
+        "LOW": "🟢 **LOW**",
+    }.get(f["severity"], f["severity"])
+    lines.append(
+        f"| {sev_emoji} | `{inst_id}` | {f['category']} | {f['message']} |"
+        f" {f['remediation']} |"
+    )
+  lines.append("")
+  return lines
+
+
+def _render_inventory_table(report: Dict[str, Any]) -> List[str]:
+  """Renders the Filestore Instance Inventory & Compliance Status table."""
+  lines: List[str] = [
+      "## Filestore Instance Inventory & Compliance Status",
+      "",
+      (
+          "| Instance ID | Location | Tier | Capacity | Reserved CIDR | Write"
+          " IOPS | Throughput | PZI | PZS | Backups | Latest Backup |"
+      ),
+      (
+          "| :--- | :--- | :--- | :--- | :--- | :--- | :--- | :--- | :--- |"
+          " :--- | :--- |"
+      ),
+  ]
+  for i in report["instances"]:
+    pzi_str = "✅ Yes" if i["satisfies_pzi"] else "❌ No"
+    if i["tier"] in ["REGIONAL", "ENTERPRISE"]:
+      pzs_str = "✅ Yes" if i["satisfies_pzs"] else "❌ No"
+    else:
+      pzs_str = "N/A"
+    b_count_str = (
+        "🔴 0" if i["backup_count"] == 0 else f"✅ {i['backup_count']}"
+    )
+    lines.append(
+        f"| `{i['instance_id']}` | `{i['location']}` | `{i['tier']}` |"
+        f" `{i['capacity_gb']} GiB` | `{i['reserved_ip_range']}` |"
+        f" `{i['max_write_iops']}` | `{i['max_throughput']}` | {pzi_str} |"
+        f" {pzs_str} | {b_count_str} | {i['latest_backup_date']} |"
+    )
+  lines.append("")
+  return lines
+
+
+def _render_remediation_actions(report: Dict[str, Any]) -> List[str]:
+  """Renders Automated Remediation Actions (Disaster Recovery) section."""
+  unprotected = [i for i in report["instances"] if i["backup_count"] == 0]
+  if not unprotected:
+    return []
+
+  lines: List[str] = [
+      "## Automated Remediation Actions (Disaster Recovery)",
+      "",
+      (
+          "To create baseline backups for unprotected instances, execute the"
+          " following commands:"
+      ),
+      "```bash",
+  ]
+  for i in unprotected:
+    loc_flag = (
+        f"--instance-location={i['location']}"
+        if i["is_regional"]
+        else f"--instance-zone={i['location']}"
+    )
+    lines.append(
+        f'CLOUDSDK_METRICS_ENVIRONMENT="{METRICS_ENV}" \\\n'
+        f"gcloud filestore backups create {i['instance_id']}-backup-$(date"
+        " +%Y%m%d) \\\n"
+        f"    --project={report['project_id']} \\\n"
+        f"    --instance={i['instance_id']} \\\n"
+        f"    --file-share={i['file_share']} \\\n"
+        f"    {loc_flag} \\\n"
+        f"    --region={i['backup_region']}"
+    )
+  lines.append("```")
+  lines.append("")
+  lines.append(
+      "> **Confirmation Required**: Would you like me to execute these backup"
+      " creation commands for the unprotected instances? Please confirm to"
+      " proceed."
+  )
+  lines.append("")
+  return lines
+
+
+def render_markdown_report(report: Dict[str, Any]) -> str:
+  """Renders the audit report in clean GitHub-flavored markdown."""
+  lines: List[str] = [
+      f"# Filestore Comprehensive Audit Report: `{report['project_id']}`",
+      "",
+  ]
+  lines.extend(_render_scorecard(report))
+  lines.extend(_render_findings_matrix(report))
+  lines.extend(_render_inventory_table(report))
+  lines.extend(_render_remediation_actions(report))
+  return "\n".join(lines)
+
+
+def main():
+  parser = argparse.ArgumentParser(
+      description="Google Cloud Filestore Comprehensive Audit Engine."
+  )
+  parser.add_argument(
+      "--project", required=True, help="GCP Project ID to audit"
+  )
+  parser.add_argument(
+      "--instance", help="Specific Filestore Instance ID (optional)"
+  )
+  parser.add_argument(
+      "--location", default="-", help="GCP zone or region (default: '-')"
+  )
+  parser.add_argument(
+      "--stale-backup-days",
+      type=int,
+      default=DEFAULT_STALE_BACKUP_DAYS,
+      help=(
+          "Threshold in days to flag stale backups"
+          f" (default: {DEFAULT_STALE_BACKUP_DAYS})"
+      ),
+  )
+  parser.add_argument(
+      "--format",
+      choices=["markdown", "json"],
+      default="markdown",
+      help="Output format (markdown or json)",
+  )
+
+  args = parser.parse_args()
+
+  auditor = FilestoreAuditor(
+      project_id=args.project,
+      instance_id=args.instance,
+      location=args.location,
+      stale_backup_days=args.stale_backup_days,
+  )
+
+  try:
+    report = auditor.run_audit()
+  except (RuntimeError, ValueError, OSError, json.JSONDecodeError) as e:
+    err_msg = str(e)
+    if args.format == "json":
+      print(json.dumps({"error": err_msg}))
+    else:
+      print(f"**Error executing audit:** {err_msg}")
+    sys.exit(1)
+
+  if args.format == "json":
+    print(json.dumps(report, indent=2))
+  else:
+    print(render_markdown_report(report))
+
+
+if __name__ == "__main__":
+  main()