Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add UnboundCompute/security-agent-skills --skill auditing-ecs-task-metadata-boundariesgit clone --depth 1 https://github.com/UnboundCompute/security-agent-skillsWrote 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/unboundcompute/security-agent-skills/auditing-ecs-task-metadata-boundaries)<a href="https://agentmods.dev/skills/unboundcompute/security-agent-skills/auditing-ecs-task-metadata-boundaries"><img src="https://agentmods.dev/badge/skills/unboundcompute/security-agent-skills/auditing-ecs-task-metadata-boundaries/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/unboundcompute/security-agent-skills/auditing-ecs-task-metadata-boundaries"><img src="https://agentmods.dev/badge/skills/unboundcompute/security-agent-skills/auditing-ecs-task-metadata-boundaries.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00165 | $0.01935 |
| Opus 5 | $0.00082 | $0.00967 |
| Sonnet 5 | $0.00033 | $0.00387 |
| Haiku 4.5 | $0.00016 | $0.00194 |
Grade A, and why
auditing-ecs-task-metadata-boundaries 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 10d 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 — 135 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Auditing ECS task metadata boundaries: when a container reaches a credential it should not
Containerized compute hands each task a role by serving credentials from a metadata endpoint the container can call. That design puts a powerful boundary inside the task: whatever the workload can reach at the credential endpoint, it can use. Two endpoints matter. The task credential endpoint returns the task role, which is fine if that role is scoped to exactly what the workload does. The host instance metadata service, if reachable from the task, returns the node's instance role, which is usually far broader and was never meant for the workload. A server-side request flaw inside the task, or a compromised sidecar, turns either endpoint into a credential-theft primitive. You audit these by checking which endpoints the workload can reach and how tightly the returned role is scoped.
When to use
- Containerized workloads assume a task or instance role served from a metadata endpoint.
- A task can reach the host instance metadata service, not only its own task credential endpoint.
- A task runs multiple containers or sidecars that share task credentials, or has a server-side request path.
Scope check
Test metadata and credential boundaries only against clusters and accounts you own or are authorized to assess, on non-production infrastructure. A confirming request retrieves live credentials, so stay inside the authorized account and treat any retrieved credential as sensitive within scope. If you can't name the authorization, stop.
The loop
-
Establish the task's real need first. Name what this workload actually must do in the cloud account: which specific actions on which resources. This is the false-positive killer: a task role scoped to exactly that need, reachable only by the workload that needs it, is correct. Name the need, then compare the reachable credentials to it.
-
Determine which metadata endpoints the task can reach. Confirm whether the workload can reach only its task credential endpoint or also the host instance metadata service. A task that can call the instance metadata service can obtain the node role, which typically exceeds any single task's need. Check the network path and any hop-limit or endpoint restriction that should block it.
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.
- 10d ago First seen · 135 lines · 165 tokens per session scan A bc4d308c756e
auditing-ecs-task-metadata-boundaries is a skill published in the GitHub repository UnboundCompute/security-agent-skills (5 stars, last pushed 2d ago), licensed MIT. It adds 165 tokens to every session and 1,935 once invoked, about $0.0008 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 2026-08-31.
Other skills, from other repositories
cloud-security-engineer
Cloud Security Engineer role bundle for cloud security posture management, IaC review, container security, and identity governance across AWS, Azure, and GCP. Orchestrates cloud-specific posture reviews and zero trust program assessments. Auto-invoked when the user needs help with cloud security configuration, IaC…
azure-review
Performs an Azure security posture review against the CIS Microsoft Azure Foundations Benchmark v2.1.0. Auto-invoked when reviewing Azure infrastructure, Entra ID configurations, NSG rules, Defender for Cloud settings, or Key Vault access policies. Walks through all nine benchmark sections, evaluates each…
gcp-review
Performs a GCP security posture review against the CIS Google Cloud Platform Foundation Benchmark v2.0.0. Auto-invoked when reviewing GCP infrastructure, IAM bindings, VPC firewall rules, Cloud Audit Logs, or GCS bucket security. Walks through all seven benchmark sections, evaluates each recommendation, and produces a…
container-security
Performs a container and Kubernetes security review against the CIS Docker Benchmark v1.6.0, CIS Kubernetes Benchmark v1.9.0, and NIST SP 800-190. Auto-invoked when reviewing Dockerfiles, Kubernetes manifests, Helm charts, or container orchestration configurations. Evaluates image security, runtime hardening, RBAC…
aws-review
Performs an AWS security posture review against the CIS Amazon Web Services Foundations Benchmark v3.0.0. Auto-invoked when reviewing AWS infrastructure, IAM policies, S3 configurations, CloudTrail settings, VPC security groups, or RDS encryption. Walks through all five benchmark sections, evaluates each…
cloud-workload-protection
../../../cloud-infra/cloud-workload-protection/SKILL.md.