Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add skills/sumonmselim/agentguard/gcpnpx skills add SumonMSelim/agentguard --skill gcpgit clone --depth 1 https://github.com/SumonMSelim/agentguardWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/sumonmselim/agentguard/gcp)<a href="https://agentmods.dev/skills/sumonmselim/agentguard/gcp"><img src="https://agentmods.dev/badge/skills/sumonmselim/agentguard/gcp.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00035 | $0.00925 |
| Opus 5 | $0.00017 | $0.00463 |
| Sonnet 5 | $0.00007 | $0.00185 |
| Haiku 4.5 | $0.00003 | $0.00093 |
Grade A, and why
gcp scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 6d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 69 lines — stays where its author put it; the contents beside it link to each section on GitHub.
GCP
Org and project structure
- Org > Folder > Project hierarchy. Folders per environment (dev, staging, prod) or team
- Separate projects per environment. Never share prod project with other envs
- Dedicated logging/security project. Ship Audit Logs, SCC findings, and VPC Flow Logs there
- Organization Policy constraints: restrict public IPs, enforce uniform bucket access, restrict allowed regions, deny SA key creation
- Billing account per team or BU. Budget alerts per project
IAM
- Least privilege. Predefined roles before custom. Never primitive roles (
Owner,Editor) in production - Separate SA per workload and environment. No shared SAs across services
- No SA keys unless unavoidable. Use Workload Identity Federation for external systems, metadata server for GCP resources
- GKE: Workload Identity per pod. Never use default compute SA (
[email protected]) — has broad Editor permissions - Grant roles at resource level. Project-level only when scope is genuinely project-wide
- IAM Recommender enabled. Act on unused permission findings
- If keys exist: rotate regularly, audit usage, delete when unused. Never commit to git
Secrets
- Store in Secret Manager. Never in env vars, container images, Terraform state, or deployment manifests
- GKE: access via Secret Manager API or mounted volumes (Workload Identity + CSI driver)
- Auto-rotation enabled where supported
- Reference by resource name in Cloud Run and GKE specs. Never inline values
Networking
- VPC per environment. Shared VPC for multi-project setups
- Private Google Access on all subnets. No public IPs needed to reach GCP APIs
- Firewall rules: deny-all default, explicit allow only. No
0.0.0.0/0except load balancer health checks - Cloud NAT for outbound internet from private instances
- VPC Flow Logs on all subnets
- Private Service Connect for managed services (Cloud SQL, Memorystore, etc.)
- Cloud Armor on all internet-facing load balancers. WAF rules + DDoS protection enabled
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 6d ago First seen · 69 lines · 0 tokens per session scan A 4c51436cd2d2
gcp is a skill published in the GitHub repository SumonMSelim/agentguard (56 stars, last pushed 1mo ago), licensed MIT. It adds 35 tokens to every session and 925 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
tcapi
Skill to call Cloud API for Tencent Cloud (腾讯云). Used for cloud automation or resource management. 当用户需要查询、创建、管理腾讯云资源，或执行云 API 自动化操作时触发。优先使用 Octop 自带 venv 中的 tccli，凭证支持全自动 OAuth 登录。.
limrun-xcode-bazel
Build a Bazel-based iOS / macOS / Apple app on Limrun's remote build execution (RBE) instead of a local Mac, and install it on a remote iOS simulator. Use when the project is a Bazel workspace (MODULE.bazel / WORKSPACE) building rulesapple / rulesswift targets and the user wants to bazel build it or run it on a…
metago-deploy-engineer
Deployment Engineer - specializes in private/cloud/edge deployments, Helm charts, canary releases, and rollback.
metago-ops-engineer
Operations Engineer - specializes in DNS, SSL, reverse proxy, production environment, and CI/CD pipeline configuration.
metago-sre-engineer
SRE Engineer - specializes in high availability, backup, disaster recovery, chaos engineering, SLO definition, and runbook generation.
external-gitcode-ascend-atb-nnal-installer
昇腾 NPU NNAL（ATB 加速库）安装技能。依赖 cann-operator-env-config 提供 Toolkit+Kernels 环境，本技能仅负责 NNAL 包的安装、环境变量配置与验证。.