google/skills
53.0
Adequate · 17 November 2026
49.9k
lines of production code
Python
primary language
5
measurements over time
What this system is
Features
Introduce Agent Platform Model Garden Deploy skill
Adds a new skill for deploying open models or custom weights from Model Garden to Agent Platform endpoints, including undeployment and cleanup. The skill provides structured instructions for safety tiers (read-only, mutating, destructive), cost estimation via a new \calculate\_cost.py\ script, and region availability checks for first-party Gemini models. It also includes a dedicated guide for copying and deploying 1P Tuned Models across projects and regions, along with reference scripts for gcloud CLI configuration and troubleshooting quota errors.
skills/cloud/agent-platform-deploy · high confidence
Introduce GKE Workload Security skill with audit and hardening workflows
A new GKE Workload Security skill has been added to provide workflows for auditing and hardening workload-level security controls. It includes a shell script to audit cluster configurations (Workload Identity, Network Policies, Shielded Nodes, Binary Authorization, and Private Cluster status), YAML assets for default-deny Network Policies and Workload Identity test pods, and documentation covering Network Policy enforcement, gVisor sandbox isolation, Pod Security Standards, and Secret Manager CSI integration.
skills/cloud/gke-workload-security · high confidence
Introduce Gemini Enterprise Agent Platform Skill Registry skill
Added the \agent-platform-skill-registry\ skill, which enables agents to interact with the Gemini Enterprise Agent Platform Skill Registry. This includes discovering skills via semantic search, listing and retrieving specific skills and their revisions, and managing the skill lifecycle (uploading, updating, and deleting skills). The skill provides a Python script (\scripts/skill\_registry\_ops.py\) to execute these operations against the Google Cloud Aiplatform API, along with documentation for generating new skill scaffolds and monitoring long-running operations.
skills/cloud/agent-platform-skill-registry · high confidence
Introduce Google Cloud Database Onboarding skill
Added a new 'cloud-databases-onboarding' skill that guides users through discovering database requirements, recommending a Google Cloud database service based on a structured recommendation matrix, and assisting in database creation. The skill enforces a three-phase workflow: requirement discovery, recommendation analysis, and implementation via a Plan-Validate-Execute pattern that drafts Terraform or gcloud CLI provisioning code for user review. It includes a Python validation script, reference prompts, and a detailed recommendation matrix covering migrations from various source databases (e.g., MySQL, PostgreSQL, Oracle) to GCP destinations (e.g., Cloud SQL, AlloyDB, Spanner, Bigtable).
skills/cloud/cloud-databases-onboarding · high confidence
Introduce IAM Access Troubleshooter skill for Google Cloud
Adds a new Google Cloud skill that diagnoses and remediates IAM access denials through two operational modes: a Requester Flow for developers to self-activate Just-In-Time (JIT) grants or escalate tickets, and a Resolver Flow for administrators to analyze policies, create deny exemptions, and provision least-privileged roles. The skill includes reference guides for safety guardrails (requiring human-in-the-loop approval for role provisioning) and MCP server configuration, along with Python scripts to find least-privileged roles and troubleshoot errors via the Policy Troubleshooter API.
skills/cloud/iam-helper-for-troubleshooting · high confidence
Introduce agent-platform-alert-configuration skill scripts
Added a new set of Python scripts to the agent-platform-alert-configuration skill to automate the setup and validation of monitoring for Vertex AI Reasoning Engines. The \gather\_agent\_info.py\ script locates agents using Cloud Asset Inventory or direct API scanning. The \check\_telemetry.py\ script verifies that required telemetry environment variables are enabled and queries Cloud Monitoring for token usage namespaces. The \create\_online\_monitor.py\ script provisions an Online Evaluator with specific quality metrics (hallucination, final response quality, tool use quality) and configurable sampling. The \analyze\_traffic.py\ script analyzes 14-day traffic time-series data to classify agent behavior (e.g., steady, bursty, seasonal) and recommend optimal dynamic alerting algorithms. Finally, \config\_utils.py\ provides shared utilities for linting PromQL queries, validating alert policy durations, and extracting alert configurations from HCL files.
skills/cloud/agent-platform-alert-configuration/scripts · high confidence
Introduce cloud-monitoring-promql-query skill for generating and validating PromQL
A new skill, cloud-monitoring-promql-query, is available to generate valid PromQL queries from Google Cloud Monitoring metric descriptors and resource parameters. The skill enforces strict project ID resolution (blocking if unavailable), maps Cloud Monitoring metric types to PromQL structures using a provided aggregation reference, and handles resource filter discovery. It includes a Python-based validation script (validate\_promql.py) that checks syntax and enforces Cloud Monitoring-specific semantic rules, such as requiring 'monitored\_resource' filters, wrapping counter metrics with rate/increase, and applying specific state filters for memory/disk utilization metrics. Unit tests are included to verify the validator's behavior against valid and invalid query patterns.
skills/cloud/cloud-monitoring-promql-query · high confidence
Introduce cloud-trace-querying agent skill for Google Cloud Trace diagnostics
The \cloud-trace-querying\ skill is now available, enabling the agent to query Google Cloud Trace spans, filter by latency or error status, correlate traces with Cloud Logging, and diagnose latency bottlenecks. This update includes the skill definition, bash-based query scripts, and reference documentation covering OpenTelemetry and GCP semantic conventions. It also provides standardized formatting assets for presenting trace data as ASCII timelines or A2UI graph structures in the user interface.
skills/cloud/cloud-trace-querying · high confidence
New 'migrate-to-rust' skill for safe, PGO-optimized legacy code migration
Adds a new \migrate-to-rust\ skill (v1.0.0) that guides the migration of legacy codebases (C++, Fortran, Java, Python, C, Shell) to 100% Safe Rust (\\#!\[forbid(unsafe\_code)\]\). The skill enforces a structured workflow including an external code confirmation gate, a bounded self-healing verification loop (max 5 iterations), A/B deterministic seed parity checks, and a 4-stage Profile-Guided Optimization (PGO) pipeline. It includes a Python telemetry script (\scripts/telemetry.py\) for tracking migration metrics, an architectural blueprint specification, and an interactive HTML infographic renderer to visualize the migration progress and results.
skills/developers/migrate-to-rust · high confidence
New API Gateway MCP Server skill for Google Cloud
A new skill has been added to configure Google Cloud API Gateway as a remote Model Context Protocol (MCP) server, allowing REST APIs to be exposed as agent tools over HTTP. This skill provides a comprehensive guide for migrating OpenAPI 2.0 specifications to OpenAPI 3.x, annotating specs with MCP extensions, and securing tool discovery with JWTs or API keys. It includes a Python validation script to ensure compliance with API Gateway's MCP requirements, reference documentation for integrating with Google ADK and API Hub, and detailed troubleshooting steps for JSON-RPC errors and public preview limitations.
skills/cloud/api-gateway-mcp-server, skills/cloud/iam-helper-for-policy-simulator · high confidence
New Agent Platform Inference skill with safety tiers and SDK examples
Added a new skill for performing inference with Google Cloud Agent Platform GenAI models, covering first-party Gemini models, third-party OpenMaaS models (Llama, DeepSeek, Qwen), and custom endpoints. The skill introduces mandatory safety and confirmation tiers (Tier R) that require explicit user approval and parameter grounding (Project ID, Region, Model ID, SDK) before executing inference to prevent unexpected costs. It includes example scripts for the Google GenAI SDK, OpenAI SDK, and Vertex AI SDK for both Gemini and OpenMaaS models, along with a verification script to ensure environment readiness.
skills/cloud/agent-platform-inference · high confidence
New Agent Platform Model Tuning skill for fine-tuning Open and Gemini models
Users can now fine-tune Large Language Models (including Open Models like Llama, Gemma, and Qwen, as well as Gemini models) using the Agent Platform infrastructure. This new skill provides a complete workflow covering environment setup, dataset preparation (with validation split enforcement), model selection, job execution, monitoring, and cost estimation. It includes procedural guidance to prevent common errors (such as unsupported model recommendations or invalid bucket creation) and ships with helper scripts for dataset preparation, cost calculation, job cancellation, and monitoring.
skills/cloud/agent-platform-tuning · high confidence
New Application Design Center design-deploy skill for GCP infrastructure
A new skill named 'application-design-center-design-deploy' has been added to guide users through the entire lifecycle of GCP infrastructure within Application Design Center (ADC). This skill replaces the previous automated 'design\_infra' tool with an agent-controlled workflow that includes local Terraform design and validation, a shifted-left best practices assessment via the ADC plan assessment API, and structured troubleshooting for deployment failures. It provides prescriptive instructions for importing Infrastructure as Code (IaC) into the ADC registry and managing application deployments using standard \gcloud\ commands.
skills/cloud/application-design-center-design-deploy · high confidence
New Filestore log-based troubleshooting skill for diagnosing mount failures
Adds a new \google-cloud-filestore-log-troubleshooting\ skill that diagnoses and remediates GCP Filestore client mount failures, including network timeouts (\ETIMEDOUT\), permission errors (\EACCES\/\EPERM\), and configuration drift. The skill provides a structured workflow to evaluate VPC ingress firewall rules (ports 2049/111), \nfsExportOptions\ ACLs, and Cloud Audit logs, with specific guidance for GKE node subnet vs. pod IP architectures. It includes a bundled Python CLI (\quick\_diagnose.py\) for automated fleet diagnosis and remediation command generation, alongside reference documents detailing error signatures, logging queries, and network specifications.
skills/cloud/google-cloud-filestore-log-troubleshooting · high confidence
New GKE Alert Configuration skill for Terraform-based monitoring
A new skill has been added to configure alerting policies for Google Kubernetes Engine (GKE) clusters using Terraform and Google Cloud Managed Service for Prometheus. The skill provides guidelines for implementing the 4 Golden Signals (latency, errors, traffic, saturation) and cluster health metrics, specifically enforcing the use of PromQL and Multi-Window Multi-Burn-Rate SLO alerts. It includes strict guardrails to manage costs associated with \kube-state-metrics\ (KSM), requiring user permission before generating alerts that depend on Tier 2 KSM metrics, and recommends native cAdvisor metrics where possible. The package includes a Python validation script (\validate\_config.py\) to lint PromQL syntax and Terraform configurations, along with reference documents detailing prerequisites, metrics catalogs, and attribution requirements for usage tracking.
skills/cloud/gke-alert-configuration · high confidence
New GKE Cluster Autoscaler skill with advanced troubleshooting and capacity buffer guidance
Users now have access to a comprehensive GKE Cluster Autoscaler skill that provides detailed guidance on enabling, optimizing, and troubleshooting cluster autoscaling. The skill introduces support for the new CapacityBuffer CRD (Preview), allowing users to pre-warm node capacity for bursty workloads using either active (warm running nodes) or standby (suspended nodes) strategies. It includes updated best practices for ComputeClass configuration, location policies, and spot instance handling. Additionally, the skill provides new diagnostic assets: a shell script to identify scale-down blockers (such as bare pods, local storage, and PDBs) and another to live-tail cluster autoscaler visibility logs for real-time monitoring and debugging.
skills/cloud/gke-cluster-autoscaler · high confidence
New GKE ComputeClasses skill with configuration templates and best practices
A new skill for configuring, optimizing, and troubleshooting GKE ComputeClasses has been added. It provides guidance on cost optimization (Spot VMs with on-demand fallback), GPU/TPU workloads, and performance tuning, along with specific rules for handling Committed Use Discounts (CUDs), capacity quotas, and node pool auto-creation. The skill includes a set of example YAML templates for various scenarios, including balanced zonal reservations, capacity quota spillover, RBAC for editor access, dynamic RWO storage classes for stateful workloads, and specific configurations for GenAI inference, Kafka, Postgres, Redis, Spark, and TPU training.
skills/cloud/gke-compute-classes · high confidence
New GKE JobSet interruption troubleshooting skill
A new diagnostic skill for GKE JobSet interruptions has been added to the Containers category. It provides a structured workflow to diagnose JobSet restart loops, spot VM preemptions, node readiness failures, and coordinator worker crashes in AI/ML training workloads. The skill includes specific MQL and PromQL queries for monitoring restarts and node status, LQL filters for log analysis, and reference documents detailing failure signatures such as host errors and NCCL timeouts. It also enforces usage attribution for telemetry tracking via environment variables and User-Agent headers.
skills/cloud/gke-ai-troubleshooting-jobset-interruption · high confidence
New GKE application onboarding skill with hardened examples
A new 'gke-app-onboarding' skill has been added to guide users through containerizing and deploying applications to Google Kubernetes Engine for the first time. The skill provides a structured workflow covering app assessment, containerization (with multi-stage build best practices), image management via Artifact Registry, and Kubernetes manifest generation. It includes concrete, production-hardened examples for both Node.js and Go applications, featuring secure configurations such as non-root users, read-only root filesystems, dropped capabilities, and digest-pinned images, along with health check endpoints and resource limits.
skills/cloud/gke-app-onboarding · high confidence
New Gemini LiveAPI skill for bidirectional WebSocket client generation
Added a new skill in \skills/cloud/gemini-live-api\ that provides instructions and reference files for generating a Gemini Enterprise LiveAPI client service class. This skill enables users to scaffold a client that connects to the Gemini Enterprise LiveAPI websocket endpoint, handling session setup/resumption, bearer token refresh, and bidirectional streaming of \ClientMessage\/\ServerMessage\ protos. It includes reference definitions for the wire protocol, session management logic for transparent resumption, and guidance for building a demo frontend/backend service to validate the integration.
skills/cloud/gemini-live-api · high confidence
New Google Cloud Filestore NFS Browser skill for read-only file inspection
Agents and cloud engineers can now inspect, search, and read files on Google Cloud Filestore (NFS) instances without local NFS clients or root privileges. This skill provides a CLI runner (scripts/nfs\_browser.py) that supports two execution engines: a recommended Cloud Run Serverless NFS Bridge (deployed via scripts/deploy\_bridge.sh) for low-latency, scale-to-zero REST access, and a GCE Jump Host via IAP SSH fallback for environments with an existing VM. The tool offers read-only operations including directory tree exploration, filename/content search, chunked file reading, and POSIX metadata inspection, with built-in safeguards like path traversal protection, binary file detection, and context-window-safe chunking to prevent token exhaustion. Also in this area: 1 commit (0 fixes) modifying skills/cloud/google-cloud-filestore-auditing, skills/cloud/google-cloud-filestore-log-troubleshooting, skills/cloud/google-cloud-filestore-nfs-browser.
skills/cloud/google-cloud-filestore-nfs-browser · medium confidence · unverified
New cloud skills for compliance mapping and monitoring chart generation
Added two new agent skills: the Assured Workloads Frameworks Navigator, which maps customer compliance requirements to Data Boundary Frameworks and CloudControls in a Google Cloud organization using the gcloud CLI, and the Cloud Monitoring Chart Generation skill, which now fully supports parsing and emitting ListTimeSeries JSON payloads into SDUI time\_series\_filter structures. Also in this area: 1 commit (0 fixes) modifying python.
python · medium confidence · unverified
New cloud-monitoring-chart-generation skill for SDUI widget creation
Added a new agent skill that transforms PromQL or ListTimeSeries queries into valid Google Cloud Monitoring Server-Driven UI (SDUI) Widget textprotos. The skill provides a three-stage workflow: computing baseline labels, synthesizing semantic plot specifications, and assembling the final protobuf output. It supports both PromQL and ListTimeSeries (time\_series\_filter) query types, ensuring they are mutually exclusive within a single dataset, and uses UUID-based filenames to prevent collisions during parallel execution.
skills/cloud/cloud-monitoring-chart-generation · high confidence
New cloud-trace-querying agent skill for GCP observability
This change introduces a new agent skill located at skills/cloud/cloud-trace-querying that provides a suite of Python scripts for interacting with Google Cloud Trace and Cloud Logging. The skill includes capabilities to fetch entire traces, retrieve specific span details, search for traces by filter, and correlate logs with trace data. It also features utilities for generating deep links to the Google Cloud Console, resolving the active GCP project ID from environment or configuration, and managing isolated virtual environments with locking and cleanup to ensure safe, concurrent execution of these tools.
(repo-wide) · high confidence
New dbt Snowflake to BigQuery migration skill
Introduces the \dbt-sf-to-bq-translator\ skill, which automates the migration of Snowflake dbt SQL models to Standardized BigQuery SQL. The skill provides a \SKILL.md\ guide and bundled Python scripts (\bulk\_translate\_via\_gcloud.py\ and \dbt\_translator.py\) that handle dialect translation, Jinja macro masking, and post-processing fixes for known BigQuery Migration Service (BQMS) limitations—such as bracket object indexing, \JSON\_VALUE\ calls on native \STRUCT\/\ARRAY\<STRUCT\>\ columns, untyped \ARRAY\<JSON\>\ casts, ISO-8601 \TIMESTAMP\ string handling, and displaced Jinja control flow blocks.
skills/cloud/dbt-sf-to-bq-translator · high confidence
New evaluation and analysis scripts for the Agent Platform Eval Flywheel
Added a suite of Python scripts to the \skills/cloud/agent-platform-eval-flywheel\ directory to support the Agent Platform Evaluation workflow. \endpoint\_evaluation.py\ and \maas\_evaluation.py\ automate running inference and evaluation against deployed endpoints and MaaS models, respectively. \parse\_adk\_traces.py\ converts Agent Development Kit (ADK) session traces into the evaluation SDK's dataset format, while \validate\_dataset.py\ checks datasets for structural compliance with the SDK schema. For result analysis, \compare\_results.py\ diffs baseline and candidate evaluation results to detect regressions, \inspect\_results.py\ renders summary and per-case scores (with optional HTML export), and \render\_html\_report.py\ generates standalone HTML reports for evaluation results and loss-analysis clusters.
skills/cloud/agent-platform-eval-flywheel/scripts · high confidence
New skill for configuring Google Cloud Monitoring notification channels
A new skill has been added to generate Terraform HCL configurations for Google Cloud Monitoring notification channels (email, SMS, Slack, PagerDuty, Pub/Sub, and webhook) and to attach them to alerting policies. The skill includes reference documentation for each channel type, a helper script to verify Pub/Sub topic compatibility (checking for unsupported message storage policies or in-transit enforcement), and a template for sensitive credentials. It also enforces standardized usage attribution for all executed commands.
skills/cloud/cloud-monitoring-notification-channels-configuration · high confidence
New skill for migrating AI workloads to self-hosted GKE inference
A new guided skill, \google-cloud-solution-guided-gke-ai-migration\, has been added to help users migrate existing AI inference workloads (from Cloud Run, Gemini API, or Gemini Enterprise Agent Platform) to self-hosted inference on Google Kubernetes Engine (GKE). This skill provides a structured, four-phase workflow—Discovery, Solution Design, Implementation, and Validation—using native \gcloud\ and \kubectl\ commands. It includes template manifests for vLLM deployments, Custom Compute Classes (CCC) for GPU provisioning, GKE Gateway API configurations for load balancing, and model staging jobs, ensuring best practices like Workload Identity, secure secret handling, and proper VRAM sizing are followed during the migration process.
skills/cloud/google-cloud-solution-guided-gke-ai-migration · high confidence
New skill for secure n-tier serverless web apps on Google Cloud
Added a new skill that guides the design and implementation of secure, multi-tier serverless web applications using Cloud Run and Cloud SQL for PostgreSQL. The skill provides a structured workflow for requirements discovery, architecture design, and implementation, including a Terraform template (\assets/main.tf\) that enforces strict network isolation via Private Service Connect, Cloud NGFW firewall policies, and VPC-internal ingress. It supports both custom domains with Google-managed SSL certificates and self-signed certificates for sandbox testing, and includes guidance for optional features like Memorystore Redis caching, Cloud CDN, and VPC Service Controls.
skills/cloud/google-cloud-solution-n-tier-serverless-web-app · high confidence
New skill for securing Google Cloud Agent Gateway multi-agent deployments
Adds a new skill that guides users through designing, deploying, and securing Google Cloud Agent Gateway solutions with dual ingress (CLIENT\_TO\_AGENT) and egress (AGENT\_TO\_ANYWHERE) patterns. The skill provides configuration assets for Ingress/Egress gateways, Model Armor guardrails (including PII and jailbreak filters), and IAP authorization policies with CEL conditions. It also includes Terraform templates for hybrid VPN and Private Service Connect (PSC) networking, along with scripts for infrastructure deployment, security validation, and troubleshooting to ensure secure multi-agent communication.
skills/cloud/google-cloud-solution-multi-agent-security · high confidence
New skill for securing and rate-limiting Google Cloud API Gateway backends
Added the \api-gateway-backend-security-and-rate-limiting\ skill, which provides a structured workflow and reference documentation for configuring client authentication (API keys and JWTs), gateway-to-backend security (using a dedicated service account to lock down Cloud Run/Functions), and per-project rate limiting (quotas) on Google Cloud API Gateway. The skill includes a Python validation script (\check\_quota\_metrics.py\) to prevent zero-downtime issues when updating quota definitions across multiple gateways, and covers security requirements for the Public Preview Model Context Protocol (MCP) support.
skills/cloud/api-gateway-backend-security-and-rate-limiting · high confidence
New troubleshooting skill for GKE TPU v6e vbar\_control\_agent OOMs
Added a new diagnostic skill for GKE TPU v6e nodes that guides users through identifying and resolving \vbar\_control\_agent\ segfaults, out-of-memory errors, and TPU device initialization failures caused by race conditions during device resets or high-frequency metrics polling. The skill provides a structured workflow to check serial console logs for OOM kills, investigate \tpu-device-plugin\ metrics checksum corruption, and detect conflicting custom metrics collection scripts, along with a validation script to verify the associated Cloud Logging queries.
skills/cloud/gke-ai-troubleshooting-tpu-vbar-oom · high confidence
Publish Developer Device Platform basic skill
Introduces the 'developer-device-platform-basics' skill, which guides users through managing remote Android devices on the Google Cloud Developer Device Platform (DDP). This includes step-by-step instructions for authentication, listing available devices, reserving sessions (with explicit billing warnings), establishing ADB connections via a provided Python forwarder script, and managing session lifecycles (status checks, expiration updates, and cancellation). The skill also includes reference documentation for the underlying API calls and a demo script for local ADB forwarding.
skills/cloud/developer-device-platform-basics · high confidence
Repository infrastructure and skills registry initialization
This change establishes the foundational structure for the skills repository. It introduces a \.gitignore\ file to exclude local development artifacts, secrets, and IDE files, and a \.lycheeignore\ file to prevent link-checking errors on local or example URLs. It adds a \.gitmodules\ file to register and pin specific Google Cloud data-cloud plugins (such as AlloyDB, BigQuery, Cloud SQL, and Spanner) as submodules. Finally, it creates an \index.json\ file that serves as the central registry, listing the names, descriptions, and entrypoint URLs for the available agent skills.
(repo-wide) · high confidence
Behavioural changes
1 commit (0 fixes) modifying skills/cloud/api-gateway-troubleshooting
A change to existing behaviour in skills/cloud/api-gateway-troubleshooting — 1 commit, 9 files: “Add api-gateway-troubleshooting skill for Google Cloud API Gateway”.
skills/cloud/api-gateway-troubleshooting · medium confidence · unverified
1 commit (0 fixes) modifying skills/cloud/bigquery-migration-data-loader
A change to existing behaviour in skills/cloud/bigquery-migration-data-loader — 1 commit, 3 files: “Add BigQuery Migration Data Loader agent skill (\bigquery-migration-data-loader\) for provisioning BigQuery datasets, batch-loading staged Parquet files into native BigQuery tables, deploying translated views from BigQuery Migration Service translation outputs, and auditing loaded table row counts”.
(repo-wide) · medium confidence · unverified
3 commits (0 fixes) modifying skills/cloud/assuredworkloads-environment-assessor
A change to existing behaviour in skills/cloud/assuredworkloads-environment-assessor — 3 commits, 5 files: “Add standardized usage attribution to Security and SecOps agent skills”; “Ground Personnel Support Model on public framework description and align deployment guidance with Assured Workloads Frameworks Console in assuredworkloads-environment-assessor”; “Register assuredworkloads-environment-assessor 1P skill and client evaluation engine”.
(repo-wide) · medium confidence · unverified
4 commits (1 fix) modifying skills/cloud/google-cloud-filestore-auditing
A change to existing behaviour in skills/cloud/google-cloud-filestore-auditing — 4 commits (1 fix), 6 files: “Add google-cloud-filestore-auditing skill for fleet disaster recovery, security access governance, and architectural reliability audits of GCP Filestore instances”; “Standardize Google Cloud product naming in first-party Filestore skill descriptions (batch 2/5: google-cloud-filestore-auditing, google-cloud-filestore-autoscale, google-cloud-filestore-nfs-browser, google-cloud-filestore-log-troubleshooting). Brand/naming copy only — frontmatter description, headings, and intro prose; no behavioral instruction changes, no file/name changes. Bumps metadata.version patch component in each modified SKILL.md”; “Fix broken Cloud SDK installation link in google-cloud-filestore-auditing skill”; and 1 more.
(repo-wide) · medium confidence · unverified
5 commits (0 fixes) modifying skills/cloud/iam-helper-for-privileged-access-management
A change to existing behaviour in skills/cloud/iam-helper-for-privileged-access-management — 5 commits, 7 files: “Update SecOps detection engineering and PAM IAM skills for YARA-L 2.0 and GA CLI flags”; “Add standardized usage attribution to Security and SecOps agent skills”; “Enforce semantic versioning (MAJOR.MINOR.PATCH) for skills”; and 2 more.
skills/cloud/iam-helper-for-privileged-access-management · medium confidence · unverified
5 commits (2 fixes) modifying skills/cloud/cloud-monitoring-alert-configuration
A change to existing behaviour in skills/cloud/cloud-monitoring-alert-configuration — 5 commits (2 fixs), 40 files: “Add cloud-monitoring-alert-configuration skill under skills/cloud/”; “Update cloud-monitoring-alert-configuration skill to use discover\_metric\_schema.py”; “Change \tags\ format to a string in cloud-monitoring-alert-configuration metadata”; and 2 more.
skills/cloud/cloud-monitoring-alert-configuration · medium confidence · unverified
Data Cloud plugins relocated to new directory structure
The Data Cloud plugins (including AlloyDB, BigQuery, Cloud SQL, Spanner, and others) have been moved to the \plugins/cloud/data-cloud\ directory. This location now vendors the plugins as git submodules, allowing them to be discovered and installed via the central \google/skills\ repository. Users can install specific plugins using the Antigravity CLI by pointing to the new paths (e.g., \plugins/cloud/data-cloud/alloydb\), or via the marketplace manifest for Claude Code and Codex.
plugins/cloud/data-cloud · high confidence
Renamed GKE TPU monitoring skills to gke-ai-troubleshooting prefix
The GKE TPU monitoring capabilities have been renamed to follow a consistent naming convention. The skill previously known as 'gke-ai-troubleshooting-tpu-dynamic-slices-monitoring' is now identified by the name 'gke-ai-troubleshooting-tpu-dynamic-slices-monitoring' (effectively a structural rename within the skills directory) and the skill 'gke-ai-troubleshooting-tpu-metrics-monitoring' is now 'gke-ai-troubleshooting-tpu-metrics-monitoring'. This change ensures that all GKE TPU-related monitoring and troubleshooting skills are grouped under the 'gke-ai-troubleshooting' prefix, improving discoverability and organization for users managing TPU workloads on GKE. The core functionality for monitoring dynamic slices and metrics remains unchanged, but the skill identifiers used for invocation have been updated.
skills/cloud/gke-ai-troubleshooting-tpu-dynamic-slices-monitoring, skills/cloud/gke-ai-troubleshooting-tpu-metrics-monitoring · high confidence
Test coverage
2 commits adding/updating tests in skills/cloud/assuredworkloads-frameworks-navigator
Test coverage in skills/cloud/assuredworkloads-frameworks-navigator — 2 commits, 6 files: “Add standardized usage attribution to Security and SecOps agent skills”; “Add the assuredworkloads-frameworks-navigator skill. It maps a customer's compliance requirements to the Data Boundary Frameworks and CloudControls in their Google Cloud organization, using the gcloud CLI”.
skills/cloud/assuredworkloads-frameworks-navigator · high confidence · unverified
Dependencies
Introduce dependency manifests for new Agent Platform skills
This change adds Python and Node.js dependency manifests for several new agent skills, establishing the runtime environment for the Agent Platform. Specifically, it defines requirements for the alert-configuration skill (using google-cloud-monitoring and google-cloud-aiplatform), the inference skill (using google-genai and openai), the tuning skill (using google-cloud-aiplatform and datasets), and the filestore-nfs-browser skill (using fastapi and uvicorn). It also introduces a basic Node.js application manifest for the GKE app-onboarding skill and a minimal manifest for the skill-registry.
(dependencies) · high confidence
Written by cca.canine.dev from the codebase's own history, inside the signed delivery this page is composed from.
How this codebase got here
Score
- CAI 51.9 → 53.0 (+1.1)
- Rubric changed (rubric-2026.11.2 → rubric-2026.11.4) — scores are not directly comparable.
Lenses
- Code Health 78.0 → 77.3 (-0.7)
- Architecture 98.7 → 98.7 (0.0)
- Maturity 67.7 → 67.7 (0.0)
- Readiness 33.9 → 32.4 (-1.5)
- Security 81.2 → 81.2 (0.0)
- Accessibility 56.4 → 63.6 (+7.2)
- Performance 100.0 → 100.0 (0.0)
Resolved (16)
- Duplicated block (13–16 lines × 2) (skills/developers/migrate-to-rust/scripts/telemetry.py)
- Duplicated block (24 lines × 2) (skills/cloud/assured-workloads-basics/scripts/manage_workload_updates.py)
- Duplicated block (24–26 lines × 2) (skills/cloud/dbt-sf-to-bq-translator/scripts/dbt_translator.py)
- Duplicated block (32 lines × 2) (skills/cloud/assured-workloads-basics/scripts/validate_params.py)
- Duplicated block (7 lines × 2) (skills/developers/migrate-to-rust/scripts/telemetry.py)
- Duplicated block (9–12 lines × 2) (skills/cloud/agent-platform-eval-flywheel/scripts/validate_dataset.py)
- Hotspot: skills/cloud/cloud-monitoring-chart-generation/scripts/textproto_util.py (skills/cloud/cloud-monitoring-chart-generation/scripts/textproto_util.py)
- Hotspot: skills/cloud/cloud-monitoring-chart-generation/scripts/validate_chart.py (skills/cloud/cloud-monitoring-chart-generation/scripts/validate_chart.py)
- Outdated: datasets
- Outdated: google-auth
- Outdated: google-cloud-aiplatform
- Outdated: google-cloud-monitoring
- Outdated: google-cloud-storage
- Outdated: google-cloud-trace
- Outdated: google-genai
- Outdated: urllib3
New (142)
- Dependency hygiene PARTLY measured — Maven/Gradle declarations read, no dependency graph resolved
- Duplicated block (13 lines × 2) (skills/platform-devices/litert-google-tensor/scripts/benchmark_on_device.py)
- Duplicated block (17 lines × 3) (skills/cloud/google-cloud-filestore-auditing/scripts/filestore_audit.py)
- Duplicated block (5 lines × 3) (skills/platform-devices/litert-google-tensor/scripts/benchmark_on_device.py)
- Duplicated block (5 lines × 4) (skills/platform-devices/litert-google-tensor/scripts/benchmark_on_device.py)
- Duplicated block (8 lines × 2) (skills/cloud/bigquery-migration-metadata-extractor/assets/extraction_stats_template.py)
- Duplicated block (8 lines × 2) (skills/platform-devices/litert-google-tensor/scripts/compile_model.py)
- Duplicated block (8 lines × 2) (skills/platform-devices/litert-google-tensor/scripts/package_ai_pack.py)
- Duplicated block (9 lines × 2) (skills/platform-devices/litert-google-tensor/scripts/benchmark_on_device.py)
- Duplicated block (9 lines × 2) (skills/platform-devices/litert-google-tensor/scripts/benchmark_on_device.py)
- Duplicated block (9–10 lines × 2) (skills/platform-devices/litert-google-tensor/scripts/install_and_verify.py)
- Edited copies of one member (3 members, 122+ corresponding lines) (skills/cloud/managed-airflow-recommendations/scripts/workload_cpu_usage.py)
- Edited copies of one member (6 members, 16+ corresponding lines) (skills/cloud/agent-platform-alert-configuration/scripts/gather_agent_info.py)
- Edited copy of a member (11 corresponding lines) (skills/cloud/cloud-logging-finops/scripts/monitoring.py)
- Edited copy of a member (115 corresponding lines) (skills/cloud/managed-airflow-recommendations/scripts/environment_health.py)
- Edited copy of a member (14 corresponding lines) (skills/cloud/cloud-trace-querying/scripts/fetch_related_logs/src/observability_scopes.py)
- Edited copy of a member (21 corresponding lines) (skills/cloud/assured-workloads-basics/scripts/manage_workload_updates.py)
- Edited copy of a member (21 corresponding lines) (skills/platform-devices/litert-google-tensor/scripts/benchmark_on_device.py)
- Edited copy of a member (22 corresponding lines) (skills/cloud/agent-platform-alert-configuration/scripts/scan_duplicates.py)
- Edited copy of a member (22 corresponding lines) (skills/cloud/api-gateway-backend-security-and-rate-limiting/scripts/check_quota_metrics.py)
- …and 122 more
Changes since last survey
- 73 commits — 67 feature/other, 6 fixes
By area
- skills/cloud — 54 commits
- (root) — 7 commits
- skills/ads — 5 commits
- plugins/cloud — 3 commits
- skills/identity — 2 commits
- .github/workflows — 1 commit
- skills/platform-devices — 1 commit
Notable commits
- fix: Fix AlloyDB ScaNN extension and parameter names in google-cloud-solution-hybrid-search-alloydb
- fix: Fix Cloud NAT flow log BigQuery field names in google-cloud-networking-observability.
- fix: Fix broken documentation link in cloud-monitoring-alert-configuration skill.
- fix: Fix broken gcloud reference link in compute-spot-capacity-advice skill.
- fix: Fix relative script paths and frontmatter metadata formatting in cloud skills.
- fix: Use UDM grouped search aliases and fix invalid network.ip in secops-hunt IoC queries.
- change: Add Assured Workloads Data Boundary frameworks navigator and author skills enhancements to the 1P Agent Registry.
- change: Add BigQuery Migration Data Loader agent skill (bigquery-migration-data-loader) for provisioning BigQuery datasets, batch-loading staged Parquet files into native BigQuery tables, deploying translated views from BigQuery Migration Service translation outputs, and auditing loaded table row counts.
- change: Add BigQuery Migration Metadata Extractor agent skill (bigquery-migration-metadata-extractor) for extracting data warehouse catalog metadata and query logs to Cloud Storage with the Google Cloud CLI, locally or on a temporary Compute Engine VM, and summarizing the extracted archives.
- change: Add BigQuery Migration Validator agent skill (bigquery-migration-validator) for validating migrated BigQuery tables against source data warehouses across schema, aggregate, and row-level data comparisons.
- change: Add gke-statefulset-without-probes public skill.
- change: Add google-cloud-release-radar skill for tracking and summarizing official Google Cloud release notes across 368 products and 4 curated product bundles.
- change: Add api-gateway-troubleshooting skill for Google Cloud API Gateway.
- change: Add context-sharing instructions to the cloud-operator skill
- change: Add litert-google-tensor skill to compile, benchmark, and package LiteRT models for Google Tensor NPUs on Pixel.
- change: Add more ad preloading fail/warning criteria to google-mobile-ads-validate skill
- change: Add the agent-platform-autorater-customization skill, which helps users build custom evaluation metrics for their agent from its own trajectories and register them in the Agent Platform evaluation service.
- change: Add the pathways-workload skill for debugging Pathways on Cloud workloads on Cloud TPU in GKE.
- change: Add threat hunting and advisory UDM reference modules to secops-hunt
- change: Address deployment walkthrough feedback in cloud-network-insights skill.
- …and 53 more
Architecture
- Containers 0 added · 0 removed · contexts 1 added · 0 removed · edges 0 added · 0 removed
Added bounded contexts (1)
- ai_pack
Written by cca.canine.dev from the codebase's own history, inside the signed delivery this page is composed from.
Survey your own repository
google/skills was measured the same way every project in this corpus was: the same rubric, at a pinned commit, with the result published in full. Point a surveyor at a repository you know and see whether you agree with it.
About this page
- The score is its most recent published measurement, taken on 17 November 2026 at a pinned commit. It is not a live figure and does not change until the project is measured again.
- Measured at commit 56fba3e82a6b6914ad7028f9fd8a1c594479e899 — the exact code this score is about.
- Scored under rubric-2026.11.4 — the same rubric and the same method as every other entry in this index.
- Measured by Canine Code Analyzer (cca.canine.dev) using cca measure-2026.11.84+preprod-d645ca9bbedd.