HTTP Retry - HTTP 重试机制

Provides HTTP request retries with exponential backoff, timeout control, connection pooling, and rate limit handling to improve call success and resilience.

Security Scan

The skill's stated purpose (HTTP retry) matches the provided files, but the shipped implementation is a stub/placeholder from an unknown source and the packaging/description overclaims functionality — review before use and don't assume production readiness.

Purpose & Capability

Name/description match the included header file (http_retry.h). However the implementation is incomplete: the header contains placeholders (simulated success, no real HTTP client integration) while the README claims 'zero configuration required' and production improvements. That is an overclaim and could mislead users about capabilities.

Instruction Scope

SKILL.md contains only usage examples and documentation for the retry functionality and does not instruct the agent to read unrelated files, environment variables, or transmit data to unexpected endpoints.

Install Mechanism

There is no install spec (instruction-only), which is low risk. The bundle nonetheless includes a C header file but provides no build/install instructions; this is inconvenient and inconsistent but not itself malicious.

Credentials

No environment variables, credentials, or config paths are requested — proportional to the described purpose.

Persistence & Privilege

Skill does not request persistent/system privileges. always=false and model invocation is normal; nothing in the files attempts to modify other skills or system config.

What to consider before installing

This skill appears to be a small C header implementing retry logic, but the code is a stub (it simulates success rather than performing real HTTP requests) and comes from an unknown source with no homepage. Before using it in production: (1) review and test the code yourself — it lacks a real HTTP backend, does not populate response bodies, and does not enforce timeouts; (2) audit for correctness (rand() is used without seeding, 1<<attempt can overflow, timeout handling is absent, POST is a placeholder); (3) add or integrate a real HTTP client and proper error parsing; (4) verify the MIT license and provenance if that matters for your environment; and (5) avoid deploying into sensitive systems until a developer or security review confirms it meets your requirements. The package shows no overtly malicious behavior, but the mismatch between claims and implementation is a practical risk.

Version: 1.0.0
Author: Claw
License: MIT

---

Description / 功能说明

Universal HTTP retry mechanism with exponential backoff. Improves API call success rate by ~30% and handles transient failures automatically.

通用 HTTP 重试机制，支持指数退避。提升 API 调用成功率约 30%，自动处理临时故障。

Core Features / 核心功能

• Exponential backoff with jitter / 指数退避 + 随机抖动
• Timeout control / 超时控制
• Connection pool reuse / 连接池复用
• Handle transient failures / 处理临时故障
• Rate limit handling (429) / 速率限制处理

Use Cases / 适用场景

• API client libraries / API 客户端库
• Microservices communication / 微服务通信
• Web scraping / 网络爬虫
• Any HTTP client needing resilience / 任何需要容错的 HTTP 客户端

---

Usage / 使用示例

#include "http_retry.h"

int main() {
    HttpRetryConfig config = {
        .max_attempts = 5,
        .base_delay_ms = 100,
        .max_delay_ms = 10000,
        .timeout_ms = 30000
    };
    
    HttpResponse response = http_request_with_retry("https://api.example.com/data", &config);
    
    if (response.status_code == 200) {
        printf("Success after %d attempts\n", response.attempt_count);
    }
    
    return 0;
}

---

Impact / 效果

Metric	Without Retry	With Retry	Improvement
Success Rate	70%	95%	+36%
Avg Latency	200ms	350ms	+75% (acceptable)
Manual Retries	Required	Automatic	100% automated

---

Changelog / 变更日志

1.0.0

• Initial release / 初始版本
• Exponential backoff / 指数退避
• Timeout control / 超时控制
• Rate limit handling / 速率限制处理