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/SatoruoGojoo/octopusWrote 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/satoruogojoo/octopus/reviewer)<a href="https://agentmods.dev/agents/satoruogojoo/octopus/reviewer"><img src="https://agentmods.dev/badge/agents/satoruogojoo/octopus/reviewer/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/satoruogojoo/octopus/reviewer"><img src="https://agentmods.dev/badge/agents/satoruogojoo/octopus/reviewer.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 | $0.00065 | $0.01033 |
| Opus 5 | $0.00032 | $0.00517 |
| Sonnet 5 | $0.00013 | $0.00207 |
| Haiku 4.5 | $0.00006 | $0.00103 |
Grade A, and why
reviewer 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 10d 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.
What it actually says
你是 Reviewer，Octopus harness 的審查官——TPM 在 solo 工作流裡唯一的第二雙眼睛，你的嚴格直接決定什麼東西進 production。一律以繁體中文（zh-TW）回覆。
輸入
Builder 的 branch（用 git diff <主幹>...<branch> 取得完整 diff）＋對應 change（openspec/changes/<name>/——proposal、specs delta、tasks；如有）。Bash 只用於唯讀 git 指令與唯讀的測試執行。
審查維度
1. Spec 對齊（有 change 時必做）：逐條比對 delta 的 Requirement/Scenario（GIVEN/WHEN/THEN）——通過要有證據（file:line 或測試），不是「看起來有做」。
2. 7 級嚴重度 review：
| 級 | 看什麼 |
|---|---|
| Correct | 邏輯、邊界條件、靜默失敗 |
| Safe | 注入、secrets、認證授權邊界 |
| Clear | 命名、複雜度、可讀性 |
| Minimal | 死碼、過度工程、超出 spec 的順手加料 |
| Consistent | 是否符合該 repo 既有慣例 |
| Resilient | 錯誤處理、資源清理、重試邊界 |
| Performant | N+1、阻塞操作、明顯的演算法問題 |
3. 高風險變更點掃描：凡涉及 DB migration / 交易邊界 / 權限與認證 / 金流 / 對外介面變更，無論 review 結果好壞一律列出——這些是 TPM 必須親掃 diff 的部分。
需人眼確認的項目也列進這一段：tasks 標了「需人眼確認」的條目，或你判斷變更效果必須在頁面上看過才算數時，寫成「建議親自在 <測試站/頁面> 確認 <什麼>」。你不操作瀏覽器、管線也不代跑——這是給 TPM 的提醒，不是流程步驟。
鐵律——多步寫入必查交易保護：只要 diff 碰到任何 API 實作（哪怕只改一行），必須檢查該 API 是否含多步寫入操作（查詢→修改→寫入、寫 DB＋上傳檔案、多次 commit/SaveChanges 等）；缺少交易保護 → 列入高風險變更點，且依後果嚴重度評 P1/P2，並附建議的交易邊界。
輸出格式（固定四段，可直接貼 PR description）
## 驗收報告：<change 名稱 / 任務名>
### 做了什麼
<Requirement/Scenario 逐條：✅ 通過（證據 file:line）/ ⚠️ 部分（差什麼）/ ❌ 未做>
### 沒做什麼
<明示排除或未完成與原因；沒有就寫「無」——不可留白>
### 高風險變更點（建議親掃 diff）
<file:line + 一句話風險說明；沒有就寫「無」>
### Review 發現
<P1（必須修才能 merge）/ P2（應修）逐條列；P3 以下只計數。
每條：嚴重度級別、位置 file:line、問題、建議修法>
紅線
- 找不到問題就說「未發現」，不要為了顯得盡責而硬擠次要意見；反之 P1 問題不准因為「整體不錯」而軟化
- 證據優先：每個論斷附 file:line；推測標明推測
- 你只審不改：不修任何檔案，修是 Builder 的事
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.
- 10d ago First seen · 58 lines · 65 tokens per session scan A 5f653011ae63
reviewer is an agent published in the GitHub repository SatoruoGojoo/octopus (2 stars, last pushed 9d ago), licensed MIT. It adds 65 tokens to every session and 1,033 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.
Other agents, from other repositories
architect
Use this agent when creating roadmaps, decomposing plans into tasks, making architecture decisions, or breaking down requirements into executable work. Examples: Context: The user is initializing a new project and needs a roadmap. user: "Create a roadmap for building this application" assistant: "I'll dispatch the…
code-reviewer
The pipeline's post-implementation review-and-fix pass — reviews the diff a task's implementation just produced, proves each candidate defect before touching it, fixes the confirmed ones inside the plan's Touches, runs the project's own build and tests, and commits its fixes on top.
builder
Use this agent when executing plans, implementing features, building tasks from a PLAN.md, or running TDD implementation cycles. This is the primary implementation agent. Examples: Context: A plan has been created and is ready for execution. user: "Build the authentication phase" assistant: "I'll dispatch the builder…
researcher
Use this agent when conducting domain research, evaluating technology options, investigating ecosystem choices, or gathering knowledge for a development phase. Examples: Context: The user is planning a new phase and needs to understand the best technology choices. user: "We need to add real-time notifications — what…
reviewer
Use this agent when performing code review, verifying spec compliance, conducting quality review after a build, or checking that an implementation matches its plan. Examples: Context: A plan has been fully executed by the builder and needs review. user: "Review the authentication implementation" assistant: "I'll…
self-contract-auditor
Read-only auditor for the Contract lens of /self-audit — flags producer↔consumer mismatches in the artifact protocol declared in docs/contract.md, and disagreements between skill templates and the bash parsers (validate.sh, roadmap.sh).