ai-gateway

Vercel AI Gateway expert guidance. Use when configuring model routing, provider failover, cost tracking, or managing multiple AI providers through a unified API.

Quality

75%

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 ./plugins/vercel/skills/ai-gateway/SKILL.md

Quality

Content

71%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.

Highly actionable and concrete with excellent code examples, but the single-file structure works against it: everything is inlined with no progressive disclosure, and dated model/pricing details will drift. Conciseness and file organization are the main improvement areas.

Suggestions

Split bulky reference material (pricing, supported providers, audit logging, cache-control tables) into references/ files, keeping SKILL.md to setup, model slug rules, and core routing patterns with clearly signaled one-level-deep links.

Move dated, time-sensitive content ("Latest Model Availability" with the March 2026 date and per-model pricing) out of SKILL.md or into a clearly marked changelog/deprecated section so staleness doesn't penalize the evergreen guidance.

Trim promotional sections ("Key Benefits", "When to Use AI Gateway" table overlapping the decision tree) and deduplicate the 402/429 handling shown in both the rate-limiting and error-handling sections.

Dimension	Reasoning	Score
Conciseness	The body is mostly concrete API guidance with little basic-concept padding, but it is noticeably over-long for a SKILL.md: marketing-style sections ("Key Benefits", "When to Use AI Gateway"), repeated 402/429 handling snippets, and dated time-sensitive material ("GPT-5.4 (added March 5, 2026)", pricing tables) that is not isolated in an old-patterns/deprecated section. It is above the 2 anchor because nearly all content is concrete rather than padded explanation.	3 / 5
Actionability	Guidance is fully executable: copy-paste-ready TypeScript for setup, routing, caching, rate-limit and budget handling, exact bash commands ("vercel link", "vercel env pull .env.local"), a complete error-handling template switching on status codes, and a provider-choice decision tree. Specific examples cover the common cases.	5 / 5
Workflow Clarity	Sequences are clear and explicit (auth setup: vercel link → dashboard enable → vercel env pull; auth resolution priority; error-recovery patterns for 429/402/503). It falls short of 5 because the doc is reference-style rather than a validated multi-step workflow — no checkpoints exist, though none are strictly required for these non-destructive operations.	4 / 5
Progressive Disclosure	Section headers and external doc links give reasonable structure, but no bundle files exist at all and ~540 lines are inlined in SKILL.md, including pricing tables, provider lists, audit-logging details, and cache-control reference material that clearly belong in separate reference files. Structure is present but content that should be separate is inline.	3 / 5
	Total	15 / 20 Passed

Description

78%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 with an explicit trigger clause and several concrete capability keywords anchored to a specific named product. The main weakness is that the "what" half relies on the meta-phrase "expert guidance" rather than stating the skill's actions directly.

Dimension	Reasoning	Score
Specificity	The description lists several concrete capabilities ("configuring model routing, provider failover, cost tracking, or managing multiple AI providers through a unified API") beyond just naming the domain. It falls short of the 5 anchor because "expert guidance" is meta-language and coverage of capabilities is not comprehensive.	4 / 5
Completeness	Both what and when are present, with an explicit "Use when..." clause covering four concrete triggers. It does not reach 5 because the "what" ("Vercel AI Gateway expert guidance") describes the artifact's nature rather than concretely stating what it does.	4 / 5
Trigger Term Quality	Natural phrases users would say in this domain are present ("model routing", "provider failover", "cost tracking", "unified API"), giving good keyword coverage. It is not a 5 because common synonyms and variations (e.g., "model fallback", "AI provider setup", "gateway keys") are missing.	4 / 5
Distinctiveness Conflict Risk	"Vercel AI Gateway" names a specific product niche and the trigger phrases (routing, failover, unified API) are gateway-specific, giving a clear niche with minimal conflict risk against generic provider or SDK skills.	5 / 5
	Total	17 / 20 Passed

Validation

81%

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 — 13 / 16 Passed

Validation for skill structure

Criteria	Description	Result
skill_md_line_count	SKILL.md is long (563 lines); consider splitting into references/ and linking	Warning
metadata_version	'metadata.version' is missing	Warning
metadata_field	'metadata' should map string keys to string values	Warning

	Total	13 / 16 Passed

Repository: openai/plugins
Path: plugins/vercel/skills/ai-gateway/SKILL.md
Commit: 5fd93af

Reviewed: 2 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.