ai-code-mode

LLM-generated TypeScript execution in sandboxed environments: createCodeModeTool() with isolate drivers (createNodeIsolateDriver, createQuickJSIsolateDriver, createCloudflareIsolateDriver), codeModeWithSkills() for persistent skill libraries, trust strategies, skill storage (FileSystem, LocalStorage, InMemory, Mongo), client-side execution progress via code_mode:* custom events in useChat.

Quality

62%

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 ./packages/ai-code-mode/skills/ai-code-mode/SKILL.md

Quality

Content

65%

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

The content is highly actionable with complete, executable code and clear guardrails, and it respects Claude's intelligence by skipping basic explanations. Its weaknesses are length/ducplication and a monolithic structure with no progressive disclosure into reference files, plus missing explicit validation checkpoints for risky code execution.

Suggestions

Collapse the duplicate Setup examples or mark one as the canonical path to reduce token cost, and trim the full React component to the essential onCustomEvent handler.

Split the client-side event-handling reference and the lazy-tools deep dive into separate one-level-deep reference files (e.g., EVENTS.md, LAZY-TOOLS.md) and link to them from a concise overview.

Add explicit validation/feedback-loop checkpoints (e.g., probe isolated-vm compatibility and verify generated code against the target driver before relying on it) to lift workflow clarity for this risky execution context.

Dimension	Reasoning	Score
Conciseness	The body avoids explaining basics Claude already knows and is mostly efficient, but it duplicates setup (createCodeModeTool then the 'recommended' createCodeMode) and embeds a 60+ line React event component, so it could be tightened rather than earning the lean 3-anchor.	2 / 3
Actionability	Provides fully executable, copy-paste-ready TypeScript with imports and concrete config, plus event/config tables and Wrong/Right Common Mistakes examples, matching the 3-anchor for specific, executable guidance.	3 / 3
Workflow Clarity	Core Patterns are numbered (1-4) and Common Mistakes act as guardrails, but there are no explicit validation/feedback-loop checkpoints for the risky operation of executing LLM-generated code, which caps workflow clarity at 2 per the guidelines.	2 / 3
Progressive Disclosure	Sections are well-organized but the file is a ~490-line monolith with no reference/asset bundle files; sizable sub-topics (client-side event handling, the lazy-tools deep dive) are inline rather than split into one-level-deep references, fitting the 2-anchor.	2 / 3
	Total	9 / 12 Passed

Description

60%

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 highly specific about the API surface and capabilities but reads as a feature/API list rather than user-facing trigger guidance. It states what the skill does well but omits any 'when to use' clause, capping completeness and weakening trigger-term quality.

Suggestions

Add an explicit 'Use when…' trigger clause naming natural user phrases (e.g., 'Use when you need the LLM to generate and run TypeScript in a sandbox, manage persistent skill libraries, or stream sandbox execution progress to the client').

Lead with natural trigger terms users would say ('sandboxed code execution', 'running generated TypeScript', 'code mode') before the API-name list, to improve trigger-term quality.

Trim or relocate the dense enumerations of driver/storage names so the opening sentence signals a distinct, non-overlapping niche.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete capabilities and named APIs ('createCodeModeTool()', 'createNodeIsolateDriver', 'createQuickJSIsolateDriver', 'createCloudflareIsolateDriver', 'codeModeWithSkills()', four storage backends, 'code_mode:* custom events'), matching the anchor for listing several specific concrete actions.	3 / 3
Completeness	It clearly answers 'what does this do' but provides no 'Use when…' clause or equivalent explicit trigger guidance, which per the judging guidelines caps completeness at 2.	2 / 3
Trigger Term Quality	Contains some natural terms ('LLM-generated TypeScript execution', 'sandboxed environments', 'skill libraries') but is dominated by API-name jargon (driver function names, 'code_mode:*') and lacks common user-phrase variations, so it falls short of the 3-anchor's broad coverage.	2 / 3
Distinctiveness Conflict Risk	The sandboxed-TS-execution niche is fairly specific, but it overlaps with general code-execution and tool-calling skills (it even cross-references ai-core/tool-calling) and lacks explicit triggers that would delimit it, so it does not reach the 3-anchor's clear non-conflicting niche.	2 / 3
	Total	9 / 12 Passed

Validation

87%

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 — 14 / 16 Passed

Validation for skill structure

Criteria	Description	Result
skill_md_line_count	SKILL.md is long (512 lines); consider splitting into references/ and linking	Warning
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	14 / 16 Passed

Repository: TanStack/ai
Path: packages/ai-code-mode/skills/ai-code-mode/SKILL.md
Commit: 4ab149f

Reviewed: about 14 hours 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.