How do I install the API Fetch Wrapper skill?

Run npx skillmds@latest add aibot88/api-fetch-wrapper in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is API Fetch Wrapper safe to use?

This skill has not completed SkillMD's automated safety review yet. Independent scanners report: SkillSpector: PASS, Skill Scanner: PASS. Capability flags: makes network calls, reads secrets. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with API Fetch Wrapper?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is API Fetch Wrapper free to use?

Yes. Installing skills from SkillMD is free. This skill is licensed under MIT.

Who published API Fetch Wrapper?

aibot88 (@aibot88) published this skill. Their other Agent Skills are listed on their SkillMD profile.

API Fetch Wrapper by aibot88 — Agent Skill

Name: API Fetch Wrapper
Rating: 4.2 (5 reviews)
Author: aibot88

API Fetch Wrapper

Wrap a public HTTP API (Open-Meteo weather as the demo) with credential handling, error normalisation, and a single retry on transient network failures. Demonstrates the production-shaped baseline for any "skill that calls an external service" — env-based secrets, structured error output, no leaked API keys in logs, and a deliberate retry policy. The Open-Meteo endpoint used here is keyless on purpose so the example runs without setup; replace `OPEN_METEO_URL` with your own host and add `process.env.MY_API_KEY` for an auth'd version.

aibot88 Updated Sep 9, 2026 3 repo stars

File contents

api-fetch-wrapper

The case that breaks first in production — a skill that hits an external service. This example covers the four things every such skill needs: secret handling, retry policy, error normalisation, and no-leak-on-failure logging.

Contract

Input (stdin, JSON):

{ "latitude": 52.52, "longitude": 13.41 }

Output (stdout, JSON):

{
  "temperatureC": 18.4,
  "windSpeedKmh": 12.6,
  "fetchedAt": "2026-05-19T08:00:00.000Z"
}

Errors — written to stderr as {"error": "...", "cause": "..."} and exit code 1. Error messages NEVER include the raw upstream response body (it might echo a secret); they include the upstream status code and a short canonical reason.

Required environment

Var	Purpose
`OPEN_METEO_URL`	Optional override of the upstream host. Defaults to `https://api.open-meteo.com`.
(none for auth)	Open-Meteo is keyless. For an auth'd API, the same skeleton reads `process.env.YOUR_API_KEY` and passes it via `Authorization` header.

Run locally

cd examples/api-fetch-wrapper
bun install
echo '{"latitude":52.52,"longitude":13.41}' | bun run src/index.ts

Adapt this

Different API — replace the URL, query params, and response shape. The retry / error-mapping skeleton stays.
API-key auth — read process.env.YOUR_API_KEY once at the top, pass it via Authorization: Bearer ${apiKey}, and ensure the key never appears in error messages.
Stronger retry — bump MAX_ATTEMPTS, add exponential backoff with jitter, or distinguish 5xx (retry) vs 4xx (do not retry) explicitly.
Cache — wrap the fetch in a TTL cache when the upstream rate-limits.

---name: api-fetch-wrapperdescription: Wrap a public HTTP API (Open-Meteo weather as the demo) with credential handling, error normalisation, and a single retry on transient network failures. Demonstrates the production-shaped baseline for any "skill that calls an external service" — env-based secrets, structured error output, no leaked API keys in logs, and a deliberate retry policy. The Open-Meteo endpoint used here is keyless on purpose so the example runs without setup; replace `OPEN_METEO_URL` with your own host and add `process.env.MY_API_KEY` for an auth'd version.license: MIT--- # api-fetch-wrapper The case that breaks first in production — a skill that hits an external service. This example covers the four things every such skill needs: secret handling, retry policy, error normalisation, and no-leak-on-failure logging. ## Contract **Input** (stdin, JSON): ```json { "latitude": 52.52, "longitude": 13.41 } ``` **Output** (stdout, JSON): ```json { "temperatureC": 18.4, "windSpeedKmh": 12.6, "fetchedAt": "2026-05-19T08:00:00.000Z" } ``` **Errors** — written to stderr as `{"error": "...", "cause": "..."}` and exit code `1`. Error messages NEVER include the raw upstream response body (it might echo a secret); they include the upstream status code and a short canonical reason. ## Required environment | Var | Purpose | |---|---| | `OPEN_METEO_URL` | Optional override of the upstream host. Defaults to `https://api.open-meteo.com`. | | *(none for auth)* | Open-Meteo is keyless. For an auth'd API, the same skeleton reads `process.env.YOUR_API_KEY` and passes it via `Authorization` header. | ## Run locally ```bash cd examples/api-fetch-wrapper bun install echo '{"latitude":52.52,"longitude":13.41}' | bun run src/index.ts ``` ## Adapt this - **Different API** — replace the URL, query params, and response shape. The retry / error-mapping skeleton stays. - **API-key auth** — read `process.env.YOUR_API_KEY` once at the top, pass it via `Authorization: Bearer ${apiKey}`, and ensure the key never appears in error messages. - **Stronger retry** — bump `MAX_ATTEMPTS`, add exponential backoff with jitter, or distinguish 5xx (retry) vs 4xx (do not retry) explicitly. - **Cache** — wrap the fetch in a TTL cache when the upstream rate-limits.

Frequently asked questions