docker-sandbox

Create, manage, and execute agent tools (claude, codex) inside Docker sandboxes for isolated code execution. Use when running agent loops, spawning tool subprocesses, or any task requiring process isolation. Triggers on "sandbox", "isolated execution", "docker sandbox", "safe agent execution", or when working on agent loop infrastructure.

Quality

78%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Passed

No findings from the security scan

Fix and improve this skill with Tessl

tessl review fix ./skills/docker-sandbox/SKILL.md

Quality

Content

65%Weight 40%Scale 1-5

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

The body is admirably lean and states clear policies and boundaries, but it operates at the level of governance rules rather than executable guidance: no commands, no explicit step ordering, and no section structure. Adding a minimal scaffold/launch command sequence with headers would lift the weakest dimensions.

Suggestions

Add a short numbered workflow (resolve workload type, inspect directory and Docker capability, create sandbox, verify service, report) so the implicit sequence becomes an explicit one with the existing verification checkpoints attached to the right steps.

Include at least one concrete executable example (e.g., the docker run invocation pattern with mounted project and published port) so actionability reaches the 'mostly executable' level instead of deferring entirely to "Use live help for the selected tool".

Add section headers (e.g., "Resolve workload", "Scaffold or launch", "Auth and isolation", "Ports and cleanup") so the short body is scannable and meets the well-organized-sections bar for progressive disclosure.

Dimension	Reasoning	Score
Conciseness	The body is ~180 words of pure directive prose with zero padding or explanation of concepts Claude already knows; every sentence carries a rule or boundary, matching anchor 5's "every token earns its place".	5 / 5
Actionability	Concrete decision rules ("use the installed `sandbox-computer` skill" vs "use `workflow-rig`") and specific auth prohibitions exist, but there are no executable commands for actually creating, launching, or verifying a sandbox — "Use live help for the selected tool" defers the key steps, matching anchor 3's "some concrete guidance but incomplete".	3 / 5
Workflow Clarity	A sequence is discernible (resolve workload type, inspect, scaffold/launch, verify) and checkpoints exist ("Verify the expected service before reporting a URL"), but the steps are scattered across thematic prose paragraphs with no explicit ordering or feedback loop, matching anchor 3 rather than anchor 4's clear sequence.	3 / 5
Progressive Disclosure	The body has no section headers and no signposted references, so while nothing needs splitting into separate files at this length, it lacks the "well-organized sections" the under-50-lines exception requires for a 5 — matching anchor 3's "some structure but could be better organized".	3 / 5
	Total	14 / 20 Passed

Description

83%Weight 40%Scale 1-5

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

A strong description that explicitly covers what the skill does and when to use it, with natural trigger terms and a clearly demarcated niche. Minor improvements possible by enumerating what "manage" entails and adding common trigger synonyms like "container" or "Dockerfile".

Dimension	Reasoning	Score
Specificity	"Create, manage, and execute agent tools (claude, codex) inside Docker sandboxes" names the domain and several concrete actions, but "manage" is generic and the operational coverage has minor gaps, matching anchor 4 rather than the comprehensive coverage of anchor 5.	4 / 5
Completeness	It clearly answers "what" (create, manage, and execute agent tools in Docker sandboxes) and "when" via an explicit "Use when..." clause plus concrete "Triggers on" phrases, exactly matching anchor 5.	5 / 5
Trigger Term Quality	Natural triggers "sandbox", "isolated execution", "docker sandbox", "safe agent execution" give good keyword coverage users would actually say, but common variations like "container" or "Dockerfile" are missing, so it falls short of anchor 5's comprehensive synonym/extension coverage.	4 / 5
Distinctiveness Conflict Risk	The Docker-sandboxed agent-tool niche is distinct with specific triggers, but the bare term "sandbox" creates minor overlap risk with other isolation or sandboxing skills, matching anchor 4 rather than minimal-conflict anchor 5.	4 / 5
	Total	17 / 20 Passed

Validation

93%

Warnings & errors only

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 15 / 16 Passed

Validation for skill structure

Criteria	Description	Result
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	15 / 16 Passed

Repository: joelhooks/joelclaw
Path: skills/docker-sandbox/SKILL.md
Commit: aa0116c

Reviewed: 8 days ago

Table of Contents

Discovery Implementation Validation

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.