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 TheArchitectit/agent-guardrails-template --skill sandbox-isolationgit clone --depth 1 https://github.com/TheArchitectit/agent-guardrails-templateWrote 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/thearchitectit/agent-guardrails-template/sandbox-isolation)<a href="https://agentmods.dev/skills/thearchitectit/agent-guardrails-template/sandbox-isolation"><img src="https://agentmods.dev/badge/skills/thearchitectit/agent-guardrails-template/sandbox-isolation/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/thearchitectit/agent-guardrails-template/sandbox-isolation"><img src="https://agentmods.dev/badge/skills/thearchitectit/agent-guardrails-template/sandbox-isolation.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 1 finding, up to medium
These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →
- medium Excessive Agency · line 63 Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.Fix: Add human-in-the-loop confirmation for destructive, irreversible, or high-impact operations. Never auto-execute commands that modify files, send data, or alter system state.
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 | $0.00019 | $0.00617 |
| Opus 5 | $0.00010 | $0.00309 |
| Sonnet 5 | $0.00004 | $0.00123 |
| Haiku 4.5 | $0.00002 | $0.00062 |
Grade A, and why
Sandbox Isolation 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 — 76 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Sandbox Isolation
The pi-guardrails extension provides a Docker-based sandbox for executing commands that might be risky, untrusted, or potentially destructive. Sandboxed commands run with network isolation, memory limits, and CPU constraints.
When To Use the Sandbox
Use the sandbox when:
- Running untrusted scripts from external sources
- Executing build commands in unfamiliar projects
- Testing destructive operations (file deletion, schema changes)
- Running user-provided code snippets
- Any command where the blast radius is unclear
Do NOT use the sandbox when:
- Running standard
git,read, oreditoperations - The command is well-understood and low-risk
- Docker is not available (check
isAvailablefirst)
How It Works
The sandbox wraps docker run with:
- No network access by default (
--network=none) - Read-only mounts for source code inspection
- Read-write mounts only for explicitly designated output paths
- Memory limits (e.g.,
512m) to prevent resource exhaustion - CPU limits (e.g.,
1.0core) to prevent CPU abuse - Timeouts (default 30s) to prevent hanging processes
- Auto-cleanup (
--rm) — containers are removed after execution
MCP Bridge
When connected, use guardrail_mcp with action sandbox_run:
action: sandbox_run
command: ["npm", "test"]
readOnlyMounts: ["/project/src"]
readWriteMounts: ["/project/test-results"]
networkAccess: false
memoryLimit: "512m"
timeout: 60000
The result includes exitCode, stdout, stderr, and timedOut fields.
Fallback When Docker Is Unavailable
If Docker is not installed or the daemon is not running, sandbox_run will return an error. In this case:
- Inform the user that sandbox is unavailable
- Ask for explicit permission before running the command outside the sandbox
- Proceed only after approval — do not assume consent
What To Do When a Sandbox Run Fails
- Exit code != 0: The command itself failed — investigate stderr
- timedOut: true: The command exceeded the timeout — increase or simplify the command
- exit code null: Docker failed to start — check Docker availability
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 · 76 lines · 19 tokens per session scan A 0315b4eceaa9
Sandbox Isolation is a skill published in the GitHub repository TheArchitectit/agent-guardrails-template (79 stars, last pushed yesterday), licensed BSD-3-Clause. It adds 19 tokens to every session and 617 once invoked, about $0.0001 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
construction-expert
Expert-level construction management, project planning, BIM, safety compliance, and construction technology. Use when the user mentions BIM, project management, safety, or building, or when the task involves Construction Management, Technologies, Standards and Regulations, or Cost Control.
kubernetes-expert
Expert-level Kubernetes cluster management, deployment strategies, networking, and production operations. Use when the user mentions containers, orchestration, devops, or cloud native, or when the task involves Kubernetes Architecture, Pods, Deployments, or Services.
docker-expert
Expert-level Docker containerization, image optimization, and container orchestration. Use this skill for building efficient Docker images, managing containers, and implementing Docker best practices.
deployment-patterns
Deployment workflows, CI/CD pipeline patterns, Docker containerization, health checks, rollback strategies, and production readiness checklists for web applications. Use when setting up deployment infrastructure or planning releases.
pydantic-ai
Build production-ready AI agents with PydanticAI — type-safe tool use, structured outputs, dependency injection, and multi-model support.
edgeone skill scanner
A local static scanner that checks agent-skill files for security risks before they are installed or used. Static analysis examines files without running them.