hosted-agents

This skill should be used when designing hosted or background agent infrastructure: sandboxed execution, remote coding environments, warm pools, session persistence, multiplayer collaboration, self-spawning agents, or Modal-style sandboxes.

Quality

56%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Low

Low-risk findings worth noting

Fix and improve this skill with Tessl

tessl review fix ./skills/hosted-agents/SKILL.md

Quality

Content

46%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 skill body is well-structured and well-organized with sensible progressive disclosure to a single reference file, but it reads as a design-philosophy essay rather than executable guidance: it is verbose in places and offers almost no concrete code, commands, or validation loops the reader can run.

Suggestions

Add concrete, executable artifacts: example Dockerfile/image-build snippets, a warm-pool sizing formula or config, and a sample git-config shell snippet, instead of "Configure git identity explicitly in every sandbox".

Insert explicit validation checkpoints into the destructive/batch flows (image rebuild -> verify build succeeded -> alert on failure; sandbox teardown -> snapshot -> confirm artifacts extracted before recycle) so the workflows include feedback loops.

Trim repetition between Core Concepts, Detailed Topics, Guidelines, and Gotchas (warm pools, server-first, user-authored commits each appear 2-3 times) and push the deeper API/schema detail into the existing reference file.

Dimension	Reasoning	Score
Conciseness	The body is mostly efficient prose but repeats the same "because" rationales and restates concepts (e.g., warm pools, server-first, user-authored commits) across Core Concepts, Detailed Topics, Guidelines, and Gotchas; tightening the redundant sections would lift it toward the efficient anchor.	3 / 5
Actionability	Guidance is almost entirely abstract design directives ("Build tools that allow agents to spawn new sessions", "Maintain a pool of pre-warmed sandboxes") with only one pseudocode-style text example and no executable commands, schema, or config the reader could apply directly, matching the high-level-hints anchor.	2 / 5
Workflow Clarity	Sequences are present (notably the numbered sandbox-to-API PR flow and the Example 1 lifecycle), but destructive/batch operations like image rebuilds, sandbox recycling, and PR creation lack explicit validate-fix-retry checkpoints, so the feedback-loop cap keeps this at 3.	3 / 5
Progressive Disclosure	The SKILL.md is an organized overview with a real one-level-deep reference (./references/infrastructure-patterns.md) that is clearly signaled with a "Read when" hint, plus well-structured external links; minor gaps (the inline Detailed Topics could push more into the reference file) keep it just below 5.	4 / 5
	Total	12 / 20 Passed

Description

67%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.

The description is a clear third-person statement covering what the skill does and when to use it, with good specificity and a distinct infrastructure niche. Its main weakness is trigger-term quality: it leans on engineering jargon rather than the phrases users would naturally say.

Suggestions

Rewrite trigger phrases in a user's natural voice (e.g., "Use when the user asks to run agents in the cloud, build a background coding agent, or scale agent execution beyond their laptop") instead of relying on terms like "warm pools" and "Modal-style sandboxes".

Add a few common synonyms/file-or-product terms a user might actually mention (e.g., "remote sandboxes", "cloud agent runners", "background agents") to broaden natural keyword coverage.

Tighten the "when" clause so it reads as an explicit trigger sentence rather than a single colon-delimited list.

Dimension	Reasoning	Score
Specificity	The description names the domain and lists several concrete capabilities ("sandboxed execution, remote coding environments, warm pools, session persistence, multiplayer collaboration, self-spawning agents, or Modal-style sandboxes"), giving good coverage with only minor gaps versus the comprehensive 5-anchor.	4 / 5
Completeness	It states both the what (designing hosted/background agent infrastructure with the listed capabilities) and the when ("This skill should be used when..."), so both halves are present though the trigger phrasing could be more concrete.	4 / 5
Trigger Term Quality	It uses mostly technical/insider phrasing ("hosted or background agent infrastructure", "warm pools", "Modal-style sandboxes") rather than the natural terms a user would actually say; common user-facing synonyms are missing, matching the "some relevant keywords but missing common variations" anchor.	3 / 5
Distinctiveness Conflict Risk	The infrastructure niche is fairly specific and the enumerated triggers carve out a clear scope, with only minor overlap risk against adjacent agent-design skills; it is mostly distinct rather than a clear niche with minimal conflict.	4 / 5
	Total	15 / 20 Passed

Validation

100%

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 — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository: muratcankoylan/Agent-Skills-for-Context-Engineering
Path: skills/hosted-agents/SKILL.md
Commit: 6dbe1a1

Reviewed: 19 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.