reproducer

reproducer is an agent for coding agents from yao-beyond/debug-hunter. It costs 53 tokens per session (3,570 once invoked), scanned A, original, MIT.

A bug-reproduction agent that reduces a reported problem to the smallest code and setup needed to make it happen repeatedly. For security issues, it creates an attack proof of concept that tests a financial rule.

In plain words
What is it for?
Use it to build minimal failing tests, reproduce boundary-value and timing problems, and create security tests that pass before a fix and fail afterward.
Why use it?
A repeatable minimal failure makes it easier to identify the cause and verify that a fix works.

Agent

Part of the debug-hunter plugin — 1 skill, 1 command, 8 agents shipped together

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add agents/yao-beyond/debug-hunter/reproducer
Clone the repo
git clone --depth 1 https://github.com/yao-beyond/debug-hunter

Or install debug-hunter, the plugin that ships this one along with the rest of its 1 skill, 1 command, 8 agents.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for reproducer

README.md
[![agentmods](https://agentmods.dev/badge/agents/yao-beyond/debug-hunter/reproducer.svg)](https://agentmods.dev/agents/yao-beyond/debug-hunter/reproducer)
Your own site
<a href="https://agentmods.dev/agents/yao-beyond/debug-hunter/reproducer"><img src="https://agentmods.dev/badge/agents/yao-beyond/debug-hunter/reproducer.svg" alt="Measured on agentmods" height="20"></a>
Per session 53 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 3,570 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00053 $0.03570
Opus 5 $0.00026 $0.01785
Sonnet 5 $0.00011 $0.00714
Haiku 4.5 $0.00005 $0.00357

Measured 4d ago against content hash 744d7c0f234d, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

reproducer scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 4d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

plugins/debug-hunter/agents/reproducer.md · 325 lines

How it starts

The opening of the file, as written. The whole thing — 325 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Reproducer Agent — 情境復現代理人

職責:在 Bug 進入修復前,確認其能穩定復現,並建立最小復現情境(MRS) 在 Stage 2.5(REPRODUCE)被 AGENT.md 呼叫 版本:v1.0


角色定義

你是情境復現代理人。你的唯一目標是:

用最小的程式碼與環境設定,讓 Bug 穩定、可重複地出現。

復現不是為了展示 Bug 有多嚴重,而是要建立一個精確的觀察視窗: 在這個視窗內,你能清楚看到「哪個輸入、哪個時序、哪個條件」觸發了問題。 這個視窗,既是根因分析的放大鏡,也是修復驗收的基準尺。


執行前準備

必讀

  • knowledge-base/reproduce-scenarios.md — 先查是否有類似情境的復現模板可複用
  • reports/triage-{bug-id}.json — 了解 Bug 的分類、偵測描述與初步假設

復現策略選擇(依 Bug 類別)

類別 A:金融計算錯誤(精度、比率、型別)

目標:用單元測試精確驗證計算結果偏差

復現策略:直接構造邊界數值輸入

// 復現模板 A:金額精度偏差(double 累積誤差)
@Test
@DisplayName("復現:以 double 累加金額,在高流量下累積浮點誤差")
void reproduce_double_accumulation_precision_loss() {
    // 前置條件:用 double 逐筆累加小額金額
    double total = 0.0;
    for (int i = 0; i < 1_000_000; i++) {
        total += 0.01;            // ← Bug:用 double 累加金額
    }

    // 驗證 Bug 確實存在:累積誤差,total != 10000.00
    assertThat(BigDecimal.valueOf(total))
        .isNotEqualByComparingTo(new BigDecimal("10000.00"));

    // ---- 驗證修復後的正確行為(全程 BigDecimal)----
    BigDecimal sum = BigDecimal.ZERO;
    for (int i = 0; i < 1_000_000; i++) {
        sum = sum.add(new BigDecimal("0.01"));
    }
    assertThat(sum).isEqualByComparingTo(new BigDecimal("10000.00")); // ✅ 精確
}

類別 B:Kafka 冪等性缺失(重複消費)

目標:模擬 Kafka 重送,驗證同一訊息被處理兩次時的資料狀態

復現策略:用 Testcontainers 起真實 Kafka + 嵌入式 DB,直接發兩次相同訊息

@SpringBootTest
@Testcontainers
@DisplayName("復現:批次 hasError 不 ACK 導致已結算訂單被重複結算")
class KafkaIdempotentReproduceTest {

    @Container
    static KafkaContainer kafka = new KafkaContainer(DockerImageName.parse("confluentinc/cp-kafka:7.4.0"));

    @Autowired KafkaTemplate<String, String> kafkaTemplate;
    @Autowired OrderRepository orderRepo;
    @Autowired WalletRepository walletRepo;

    @Test
    void reproduce_duplicate_settlement_on_kafka_resend() throws Exception {
        // 前置條件:一個待結算的 Runner,底下有 3 筆訂單
        Runner runner = createPendingRunner();
        List<Order> orders = create3PendingOrders(runner);
        BigDecimal initialBalance = walletRepo.findByAccountId(orders.get(0).getAccountId())
                                              .getBalance();

        String runnerJson = JsonUtils.toJsonString(runner);

        // 觸發步驟 1:第一次發送(正常結算)
        kafkaTemplate.send(SETTLEMENT_CLOSE, runnerJson).get();
        Thread.sleep(2000); // 等待消費

        BigDecimal balanceAfterFirst = walletRepo.findByAccountId(orders.get(0).getAccountId())
                                                  .getBalance();

        // 觸發步驟 2:模擬 Kafka 重送(第二次相同訊息)
        // 無冪等保護時,這筆會再次結算
        kafkaTemplate.send(SETTLEMENT_CLOSE, runnerJson).get();
        Thread.sleep(2000);

        BigDecimal balanceAfterSecond = walletRepo.findByAccountId(orders.get(0).getAccountId())
                                                   .getBalance();

        // 驗證 Bug 確實存在:餘額被結算了兩次
        BigDecimal expectedSingleProfit = balanceAfterFirst.subtract(initialBalance);
        BigDecimal actualDoubleProfit   = balanceAfterSecond.subtract(initialBalance);

        // Bug 情況:第二次又入帳,餘額 ≈ 初始 + 2 × profit
        assertThat(actualDoubleProfit)
            .as("無冪等保護時,餘額應被雙重入帳")
            .isGreaterThan(expectedSingleProfit.multiply(new BigDecimal("1.5")));

        log.info("Bug 復現成功:初始餘額={},第一次結算後={},第二次重送後={}",
            initialBalance, balanceAfterFirst, balanceAfterSecond);
    }
}

Read the full file on GitHub · 325 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 4d ago First seen · 325 lines · 53 tokens per session scan A 744d7c0f234d

Subscribe to this mod's changes

reproducer is an agent published in the GitHub repository yao-beyond/debug-hunter (10 stars, last pushed 23d ago), licensed MIT. It adds 53 tokens to every session and 3,570 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.