CLAUDE.md@docs/measurements/2026-10-24-s5-outputsA
CLAUDE.md@docs/measurements/2026-10-24-s5-outputs from kopp0510/claude-dd is a CLAUDE.md file: the standing instructions Claude Code reads in that repo.
mdr add kopp0510/claude-dd/docs/measurements/2026-10-24-s5-outputs/CLAUDE.md@git:20261024.ce3dd5dPins git:20261024.ce3dd5d in mdr.lock. Pin the fingerprint instead.
- Audit grade
- A, 14 of 14 checks passed
- Kind
- CLAUDE.md file
- Repository
- kopp0510/claude-dd
- Path
docs/measurements/2026-10-24-s5-outputs/CLAUDE.md- Versions recorded
- 3
- Latest version
- git:20261024.ce3dd5d, committed Oct 24, 2026
- First version recorded
- Oct 24, 2026
- Size
- 1,771 bytes
- Repository stars
- 12
- License
- MIT
Collected from public GitHub. Every version is saved exactly as it was and never changes. Install it at an exact version with the mdr tool, and every file is checked before it reaches your agent. The main file is checked against the SHA-256 fingerprint recorded here, and the others against the fingerprints GitHub keeps for its source commit. The automatic checks below grade the latest version, and the same checks always give the same file the same grade.
What the file says
# docs/measurements/2026-10-24-s5-outputs 裁決審查 S5 量測的**結果**（輸入在旁邊的 `2026-10-24-s5-inputs/`，解讀在上一層 `2026-10-23-reviewer-real-sessions.md`「拿掉裁決者的對照」一節）。 ## 檔案 - `stats.jsonl` —— `../2026-10-24-s5-inputs/stats.py` 對 8 發的輸出，一發一列（耗時、派工前 `pre_s`、收尾 `post_s`、工具次數、費用、作廢理由） - `checks.out` —— `../2026-10-24-s5-inputs/checks.py` 的輸出（每個讀者讀 diff 檔全檔缺幾行、父報告晚於審查者、工具結果裡的 `/Users/`、跨發存取） - `sec1.py` —— 跑完才補的檢查：每個讀者對 diff 檔「第一段」（`git diff` 本體）缺幾行。`checks.py` 算的是全檔；`direct.txt` 要求讀完全檔，實際多數讀者沒讀完第二段（整檔內容），所以另外用第一段當比較基準。S4 那 12 發也一起算 - `sec1.out` —— 它對 S5 8 發與 S4 12 發的輸出（`P` 是父 session，`r` 是審查者；S4 的裁決者是 general-purpose，不列） - `stats-noadvisor.jsonl`、`checks-noadvisor.out`、`modelusage-noadvisor.out` —— B 組關掉 advisor 重跑的 4 發（`G2-dhk-3` 這種）：同兩支腳本的輸出，加上每發 result 的 `modelUsage` 鍵（確認沒有 advisor 的模型）；`sec1.out` 尾端也接上這 4 發 - `reports/` —— 每個審查者的最後一則文字（`d1`）與父 session 的最後報告（`d0.parent`），檔名 `<發代號>.d<層>.<agent id 前 6 碼>.<description>.md` ## 約束 - 這是當時的輸出，不要手改；工作目錄路徑已換成 `<s5>` - 重跑任何腳本都要 session 暫存目錄裡的快照、`logs/` 與 `~/.claude/projects/` 底下各快照的紀錄；session 清掉就重跑不了
Read the whole file at its exact version.
How to install
mdr add kopp0510/claude-dd/docs/measurements/2026-10-24-s5-outputs/CLAUDE.md@sha256:5dc555b317c6d17fThe line at the top of this page pins the label git:20261024.ce3dd5d, which is what most people want. This one pins the sha256 fingerprint of the contents instead, so it survives a label being moved. Either way the exact version is written to mdr.lock, and mdr install downloads and checks those contents again, so it installs them exactly or stops with an error.
[](https://markdownregistry.com/a/art_m4eqjdtdxwmbizvr)
2 badge image fetches in 30 days. What the badge shows.
Versions
The registry has recorded 3 versions of this file, all committed on Oct 24, 2026.
| version | committed | commit | size | audit | |
|---|---|---|---|---|---|
| git:20261024.ce3dd5d latest | 2026-10-24 | ce3dd5d | 1,771 B | A | view · diff |
| git:20261024.c0406be | 2026-10-24 | c0406be | 1,481 B | A | view · diff |
| git:20261024.223f0ba | 2026-10-24 | 223f0ba | 1,459 B | A | view |
Audit of the latest version
- pass: Size between 200 bytes and 200 KB (1771 bytes)
- pass: No zero-width or bidi control characters
- pass: No instruction hidden inside an HTML comment
- pass: No link to an exfiltration or paste host
- pass: No credential-shaped string
- pass: No instruction to send local credentials anywhere
- pass: No text hidden with inline styles
- pass: No prompt-injection phrasing
- pass: No curl or wget piped into a shell
- pass: No recursive delete of root, home or parent
- pass: No instruction to read or print local credentials
- pass: No base64 blob over 200 characters
- pass: No link to a raw IP address
- pass: No script tag
Source
kopp0510/claude-dd · 12 stars · license MIT · pushed 2026-10-30 · branch main
API
GET https://markdownregistry.com/api/v1/artifacts/art_m4eqjdtdxwmbizvr GET https://markdownregistry.com/api/v1/resolve?ref=kopp0510/claude-dd/docs/measurements/2026-10-24-s5-outputs/CLAUDE.md GET https://markdownregistry.com/api/v1/blob/5dc555b317c6d17fdc40a253e27b2f06332d7b4f5813db625629280a2ba7f50c
Your AI assistant does the legwork. You hear about the deals worth your word. Hand yours the standing instructions at modelranch.com and it joins the network that reads files like this one.
Guides: what a CLAUDE.md file is · keeping AGENTS.md and CLAUDE.md in sync · which agent file does what.