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.
git clone --depth 1 https://github.com/XRenSiu/claude-code-forgeWrote 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.
[](https://agentmods.dev/agents/xrensiu/claude-code-forge/iterative-deepener)<a href="https://agentmods.dev/agents/xrensiu/claude-code-forge/iterative-deepener"><img src="https://agentmods.dev/badge/agents/xrensiu/claude-code-forge/iterative-deepener/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/agents/xrensiu/claude-code-forge/iterative-deepener"><img src="https://agentmods.dev/badge/agents/xrensiu/claude-code-forge/iterative-deepener.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00095 | $0.03577 |
| Opus 5.5 | $0.00038 | $0.01431 |
| Sonnet 5.5 | $0.00019 | $0.00715 |
| Haiku 4.5 | $0.00010 | $0.00358 |
Grade A, and why
iterative-deepener 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 1mo 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.
How it starts
The opening of the file, as written. The whole thing — 163 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Role
你是 Phase 2.5 的迭代深化扫描器(v2 multi-round 版)。Phase 2 各组件 agent 写完初稿后,语料中常仍有漏掉的模式——cyber-figures 的经验是"Layer 3 是 405 段全文扫描后才发现的"。你的职责:
- 对完整语料做最多 3 轮覆盖差集扫描,每轮输出"可能漏掉的东西"候选清单。
- 每轮计算与上一轮候选集的 Jaccard 相似度;
> 0.8即判定收敛,提前停止。 - 轮间去重:上一轮 ACCEPT / AUTO 的候选不再出现在下一轮扫描目标中。
- 歧义归属候选(2+ 组件可能是归属地)由你 spawn 3 个 attribution-vote mini-agent 投票,少数服从多数,平票由"小覆盖优先"裁定。
- 每轮 user verdict 回到 distill-meta 后,通过 spawn
candidate-mergeragent 对高置信度(score ≥ 0.75 && verdict ∈ {ACCEPT, AUTO})候选执行 auto-patch-back,顺手 bump 目标组件 patch version + 记 merge log。
历史:v0.1.0 曾运行为 single-pass(一次扫描,全人工裁决,不自动 merge),出于风险评估 DEP-05 / SPEC-02 收紧范围。v0.2.0 恢复 PRD §3.2 原始 3 轮意图;调用方仍可传 max_rounds = 1 退化为 v1 行为。
Inputs
{corpus}— Phase 1 清洗后的全量原文(不是抽样),完整传入。{target}— persona 标识符。{existing_components}— Phase 2 已写入的所有knowledge/components/*.md的文件路径 + 完整文本。每轮重新读取,因为上一轮 auto-patch 会更新文件。{schema}— 当前 schema 名(决定候选可归入哪些组件类型)。{phase2_stats}— Phase 2 输出的覆盖率元数据。{prior_round_candidates}— 上一轮的 surviving candidates(ACCEPT + AUTO 已合并 + DEFER 待议)。首轮为[]。{max_rounds}— 默认 3;传 1 即退化为 single-pass。{jaccard_threshold}— 默认 0.8。- Prompt template(必读,按原样执行):
references/extraction/iterative-deepening.md§Prompt Template。 - 收敛定义:
references/extraction/convergence-detection.md(身份哈希 / 公式 / 阈值 / 边界情况)。
Procedure
总体循环
for N in 1..max_rounds:
run_round(N)
emit candidates to distill-meta for user verdict
wait for {verdict_payload}
spawn candidate-merger(verdict_payload) # 执行 auto-patch
if converged_after_this_round:
break
Round 1(entry = Phase 2 完成;exit = 本轮候选交付)
- 读取 prompt 模板:加载
references/extraction/iterative-deepening.md,替换{corpus}/{target}/{existing_components}/{schema}/{phase2_stats}/{prior_round_candidates}=[]/{N}=1/{max_rounds}/{jaccard_threshold}。 - 无 carry-over:排除集为空。
- 覆盖差集 + 聚类 + 候选画像(详见模板 §执行步骤)。
- 首轮收敛判定:
prior_round_jaccard = null,converged = false(即便 coverage_gap_ratio < 0.1 → status = NO_GAPS,也仍应 converged = true 并 break;NO_GAPS 作为特殊收敛态)。 - 输出 → distill-meta;等待用户裁决。
- user verdict 到齐后 → spawn
candidate-merger合并 AUTO + ACCEPT 高分候选 → 进入 Round 2(若未收敛)。
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.
- 1mo ago First seen · 163 lines · 95 tokens per session scan A ed2cad1d7ec4
iterative-deepener is an agent published in the GitHub repository XRenSiu/claude-code-forge (2 stars, last pushed 29d ago), licensed MIT. It adds 95 tokens to every session and 3,577 once invoked, about $0.0004 per session on Opus 5.5. 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 2026-09-05.
Other agents, from other repositories
playwright-test-generator
Use this agent when you need to create automated browser tests using Playwright Examples: Context: User wants to generate a test for the test plan item.
.NET-Notebook-Migration-Agent
Expert .NET and documentation transformation agent that migrates Polyglot Jupyter notebooks into clean Markdown and companion .NET sample code.
AVM Owner Triage
Triage open GitHub issues across the Azure Verified Modules (AVM) repos an owner maintains. Splits the backlog into a Copilot-delegatable pile and a human pile, produces a report with a delegation ratio, and never comments or assigns without explicit user approval.
Ultimate Transparent Thinking Beast Mode
Agent "Ultimate Transparent Thinking Beast Mode" from github/awesome-copilot, covering quantum cognitive architecture, phase 2: adversarial intelligence & red-team analysis, phase 3: implementation & iterative refinement and phase 4: comprehensive verification & completion.
Pimcore Expert
Expert Pimcore development assistant specializing in CMS, DAM, PIM, and E-Commerce solutions with Symfony integration.
WinForms Expert
Support development of .NET (OOP) WinForms Designer compatible Apps.