Getting it into your agent
This one installs as part of its plugin. Adding the marketplace and installing the plugin brings it with everything else the plugin ships.
/plugin marketplace add vinnie357/claude-skills/plugin install agent-sandboxingWrote 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/commands/vinnie357/claude-skills/build-image)<a href="https://agentmods.dev/commands/vinnie357/claude-skills/build-image"><img src="https://agentmods.dev/badge/commands/vinnie357/claude-skills/build-image/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/commands/vinnie357/claude-skills/build-image"><img src="https://agentmods.dev/badge/commands/vinnie357/claude-skills/build-image.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 | $0.00025 | $0.00731 |
| Opus 5 | $0.00013 | $0.00365 |
| Sonnet 5 | $0.00005 | $0.00146 |
| Haiku 4.5 | $0.00003 | $0.00073 |
Grade C, and why
build-image scanned grade C with 1 finding 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 5d 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.
Recursive force deletehighDestructive command
rm -rf with a variable or a broad path is one typo away from removing the wrong tree.
rm -rf $BUILD_DIR How it starts
The opening of the file, as written. The whole thing — 82 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Build the workload image that SandboxTemplates reference.
Skills to load
/agent-sandboxing:claude-code-on-sandbox/core:mise/core:anti-fabrication
Steps
1. Resolve build context
The plugin ships:
templates/Dockerfile.claude-code— mise-driven Dockerfile, non-root agent user, BuildKit cache mounts.templates/mise.toml.claude-code— pinsnode = "lts"+claude-code = "latest"via the mise registry short name. Do not switch tonpm:@anthropic-ai/claude-code— the short name is the convention here.
Stage them into a build context:
BUILD_DIR=$(mktemp -d)
cp <CLAUDE_PLUGIN_ROOT>/templates/Dockerfile.claude-code $BUILD_DIR/Dockerfile
cp <CLAUDE_PLUGIN_ROOT>/templates/mise.toml.claude-code $BUILD_DIR/mise.toml
2. Build
Pick --load for local-only iteration, --push for registry publication. Default to --load if neither is specified.
REGISTRY=${REGISTRY:-ghcr.io/$USER}
TAG=${TAG:-$(date +%Y%m%d)}
docker buildx build \
--tag ${REGISTRY}/claude-code:${TAG} \
${PUSH:+--push} ${LOAD:+--load} \
--file $BUILD_DIR/Dockerfile \
$BUILD_DIR
3. Verify
docker run --rm ${REGISTRY}/claude-code:${TAG} --version
Expected: a Claude Code version string (the image's CMD is --print --output-format json, so passing --version overrides it).
If the image returns no output or fails, inspect the Dockerfile — likely the mise install step couldn't reach the npm registry from inside the build, or the node = "lts" resolved to an unsupported version.
4. Update the SandboxTemplate
Once the image is built and pushed, edit the installed SandboxTemplate to use it:
kubectl set image sandboxtemplate/claude-code-kata claude=${REGISTRY}/claude-code:${TAG}
Or re-apply from the plugin template with placeholders substituted:
sed -e "s|REGISTRY/claude-code:TAG|${REGISTRY}/claude-code:${TAG}|" \
<CLAUDE_PLUGIN_ROOT>/templates/SandboxTemplate.kata.yaml | kubectl apply -f -
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.
- 5d ago First seen · 82 lines · 25 tokens per session scan C e1ffd1321165
build-image is a command published in the GitHub repository vinnie357/claude-skills (25 stars, last pushed 3d ago), licensed MIT. It adds 25 tokens to every session and 731 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it C with 1 finding (recursive force delete). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-03.
Other commands, from other repositories
create-docker-mcp-tunnel
Stand up an Anthropic MCP tunnel locally with Docker Compose so Claude can call a private MCP server (manual-credentials quickstart).
cisco-isovalent-platform-setup
Install the Isovalent platform on Kubernetes (Cilium CNI, Hubble observability, Tetragon eBPF runtime security) in either OSS or Enterprise edition. Renders Helm values and install/upgrade scripts; this is the platform install prerequisite for the Splunk Observability + Splunk Platform integration done by…
application-crud
A command reference for creating and managing Coolify applications connected to private GitHub repositories. It covers application records, deployment settings, and common API errors.
review-helm
Run Helm chart review for structure, best practices, and K8s security/operational issues in templates.
create-operator
Guided workflow to create a complete Kubernetes CRD operator.
docker-harden
Audits and hardens Dockerfiles for size and security.