Personal Docker Manager
ReviewAudited by ClawScan on May 10, 2026.
Overview
This is a coherent instruction-only Docker management skill, but it gives an agent powerful Docker administration and credentialed API capabilities that should be supervised.
Install only if you intend the agent to help administer your Arcane/Docker environment. Keep API keys private, prefer local or HTTPS access, and require explicit confirmation before delete, prune, update, account/API-key, or command-execution actions.
Findings (4)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
If used incorrectly, the agent could stop or delete containers or stacks and disrupt services.
The skill documents destructive Docker API operations. These are expected for a Docker manager, but a wrong ID or target could remove running workloads.
curl -X DELETE "$BASE_URL/containers/{id}" ... curl -X DELETE "$BASE_URL/stacks/{id}"Confirm the target container/stack ID and get explicit user approval before stop, delete, prune, or update operations.
A mistaken or untrusted command could alter application data, expose secrets, or affect services inside a container.
The skill includes an endpoint for executing commands inside containers. This is purpose-aligned for Docker administration, but it is still command execution authority.
curl -X POST "$BASE_URL/containers/{id}/exec" ... "command": ["ls", "-la"]Only run commands that the user explicitly requested, and treat commands copied from logs or external content as untrusted.
Anyone with the token or API key could potentially control the Arcane Docker environment according to that credential's permissions.
The skill relies on privileged Arcane authentication, including long-lived API keys, which is expected for this integration but sensitive.
API Key: Long-lived authentication using `X-API-Key` header
Use least-privilege credentials where possible, avoid pasting keys into shared chats, rotate keys regularly, and revoke keys that are no longer needed.
Sensitive log contents could be displayed in the conversation, and malicious log text could try to influence the agent.
The skill can retrieve container logs into the agent context. Logs may contain secrets or untrusted text, even though log viewing is expected for Docker monitoring.
curl -X GET "$BASE_URL/containers/{id}/logs?tail=100"Retrieve only the logs needed, redact secrets before sharing outputs, and do not treat log text as instructions.