Ai Sdk Core
v0.1.0Build backend AI with Vercel AI SDK v6 stable. Covers Output API (replaces generateObject/streamObject), speech synthesis, transcription, embeddings, MCP tools with security guidance. Includes v4→v5 migration and 15 error solutions with workarounds. Use when: implementing AI SDK v5/v6, migrating versions, troubleshooting AI_APICallError, Workers startup issues, Output API errors, Gemini caching issues, Anthropic tool errors, MCP tools, or stream resumption failures.
⭐ 2· 1.9k·1 current·1 all-time
byVeera@veeramanikandanr48
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Benign
high confidencePurpose & Capability
Name/description, README, SKILL.md and the templates all align: this is a developer guide + templates for building backend code with the Vercel AI SDK (v5/v6), multi-provider examples, and troubleshooting. The files and templates (openai/anthropic/google/workers) are expected for this purpose; no unrelated env vars or binaries are declared in metadata.
Instruction Scope
SKILL.md is a large, example-rich runtime doc. Most instructions are standard code samples. Two things to note: (1) several examples show executing a local command via MCP transport that uses 'npx -y @modelcontextprotocol/server-filesystem' (i.e., downloading and running an npm package at runtime) — this is a legitimate MCP pattern but executes third-party code and should be vetted before running in a production environment; (2) template tool implementations show file I/O (fs.readFile, fs.unlink) and payment/delete examples that require human-in-the-loop approvals — expected for agent/tool patterns but these capabilities permit reading/modifying local files if you plug the templates into a deployed agent.
Install Mechanism
There is no install spec for the skill bundle itself (instruction-only install), which is lowest-risk. The SKILL.md recommends running npm install and shows npx in examples — typical for code samples. No downloads from nonstandard URLs or embedded installers are present in the package metadata.
Credentials
The skill metadata declares no required environment variables. The documentation appropriately documents provider API keys (OPENAI_API_KEY, ANTHROPIC_API_KEY, GOOGLE_GENERATIVE_AI_API_KEY) for the example projects; this is proportional to the stated purpose. There are no unexpected secret requests in the manifest.
Persistence & Privilege
Skill flags: always:false and user-invocable:true. The skill does not request persistent privileges or attempt to modify other skills or system-wide configuration. Templates include code that would run in your project (e.g., caching, telemetry hooks) but the bundle itself does not persist credentials or enable forced inclusion.
Assessment
This skill is primarily documentation + templates for the Vercel AI SDK and appears coherent with that purpose. Before using it: (1) vet and update provider/model references — the package/docs include outdated model names and version notes (the included VERIFICATION_REPORT itself flags outdated packages and model names); (2) do not blindly run example npx commands (the MCP stdio example runs an npm package at runtime) — audit any third-party package you install or run, especially when using 'npx -y' which downloads+executes code; (3) treat example tools that read/write files or perform destructive actions as code templates only — add proper human approvals, sandboxing, and permissions before deploying; (4) supply provider API keys only in your own deployment environment (the skill does not request keys itself); (5) test templates in a local/dev environment and update dependencies (zod, provider libs) to current versions before production use.Like a lobster shell, security has layers — review code before you run it.
latestvk972yxtw7g63xaa9xqqr1xdfe9808a6m
License
MIT-0
Free to use, modify, and redistribute. No attribution required.