04-operation-jev

04-operation-jev is a skill for Claude Code, Codex from miyatti777/ai-plc. It costs 61 tokens per session (6,072 once invoked), scanned A, original, MIT.

An experimental workflow for carrying out planned tasks and storing the resulting artifacts. It adds Jev monitoring, which checks selected progress conditions and reports a one-line hint for review.

In plain words
What is it for?
Running approved tasks, saving their outputs, monitoring selected layers, and collecting unresolved Jev decisions.
Why use it?
It provides an extra signal when progress may need attention while keeping decisions with the main process and user. It also tracks whether each hint was accepted or rejected.

Skill for Claude CodeCodex

Written for no agent in particular: nothing here depends on one. Also seen: reads .claude/ paths; mentions subagents; mentions Codex.

Good fit Running approved tasks, saving their outputs, monitoring selected layers, and collecting unresolved Jev decisions.

Compare 6 skills from other repositories ↓
Install with agentmods
npx agentmods add skills/miyatti777/ai-plc/04-operation-jev
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.

Any agent
npx skills add miyatti777/ai-plc --skill 04-operation-jev
Clone the repo
git clone --depth 1 https://github.com/miyatti777/ai-plc

Made for: Claude Code, Codex.

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 04-operation-jev

README.md
[![agentmods](https://agentmods.dev/badge/skills/miyatti777/ai-plc/04-operation-jev/github.svg)](https://agentmods.dev/skills/miyatti777/ai-plc/04-operation-jev)
Your own site
<a href="https://agentmods.dev/skills/miyatti777/ai-plc/04-operation-jev"><img src="https://agentmods.dev/badge/skills/miyatti777/ai-plc/04-operation-jev/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.

agentmods 80×15 button for 04-operation-jev

Your own site · 80×15
<a href="https://agentmods.dev/skills/miyatti777/ai-plc/04-operation-jev"><img src="https://agentmods.dev/badge/skills/miyatti777/ai-plc/04-operation-jev.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 61 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 6,072 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
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 5.1 $0.00061 $0.06072
Opus 5.5 $0.00024 $0.02429
Sonnet 5.5 $0.00012 $0.01214
Haiku 4.5 $0.00006 $0.00607

Measured 16d ago against content hash fbc5844b66a8, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

04-operation-jev 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 16d 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.

experimental/jev/skills/ai-plc-jev/04-operation-jev/SKILL.md · 121 lines

How it starts

The opening of the file, as written. The whole thing — 121 lines — stays where its author put it; the contents beside it link to each section on GitHub.

AI-PLC Stage 4: Operation(Jev実験版)

🧪 実験版(ai-plc-jev 1.12.0-exp.2): 公開 core の .claude/skills/ai-plc/04-operation/SKILL.md(v2.8)の代わりに /04-operation-jev で使う試行用のスキル。本体の SKILL・rules は書き換えない。 公開 core 2.8 との違いは Jev 部分だけ: ①下の「Jev監視ルール」②Phase 5.5b / 6b の Jev 呼び出し(core の「呼ばない」段落を置き換え)③未確認の Jev 判定の件数表示と回収 ④Next Action のコピペ用プロンプトを -jev 表記に(Phase 1・4 の /03-construction の案内も -jev)⑤自動完走中(RUL_plc_session §10)の Jev の扱い(Jev監視ルールの例外)。このほかは、Phase 8 / 9-11 の参照先を本体側のパス(.claude/skills/ai-plc/04-operation/)に変えただけ。詳細は .claude/skills/ai-plc-jev/README.md。

Jev監視ルール(core の rules は Jev 監視を実験版の /01〜04-*-jev だけに限定しており〔RUL_plc_adaptive §5 実行ルール6〕、その具体的な手順を本スキルで定める): intent.yamlで jev_monitor: true のLayerでは、Phase 5.5b(BT-A)/ 6b(BT-B/C)で .claude/ai-plc-jev/scripts/jev_bt_monitor.py を呼び、判断専用モデルJevに「異常ありか」の1問だけを聞く。数え上げ条件(完了率50%・完了5件超・ad-hoc 2件以上・全完了)はスクリプトがコードで判定する。Jevの出力は1行ヒントのみで止める権限を持たず、Backtrackの種類判断と提案文はメインモデルが行い、実行はRUL_plc_adaptive §5 実行ルール1(承認後に実行)に従う。ヒントは独立checker(同 実行ルール5)の代わりにしない。Jevが使えない・送信禁止に当たる場合はスキップして通常どおり続行する。ヒントの有無にかかわらず、各判定のあとにユーザーの確認(妥当=accept/外れ=reject)を --override <decision_id> accept|reject で記録する(試行の評価に使う。decision_id は判定の出力行に出る)。未確認の判定の回収: 記録漏れはコードで拾う。進捗ダッシュボードに python3 .claude/ai-plc-jev/scripts/jev_bt_monitor.py --pending --layer <Layer> の件数を「🧭 未確認の Jev 判定: N件」として出す(jev_monitor: true の Layer だけ。0件なら 0件と出す)。Next Actionのコピペ用プロンプトはすべての選択肢に、そのコマンドの出力の「貼り付け用」1行(Jev判定 X・Y・Z は accept)を含める(記憶で書かず毎回コマンドの出力から写す。0件なら書かない)。利用者の返答に「Jev判定 X・Y・Z は accept」「全部accept」等があれば、利用者に示した貼り付け用1行の ID だけを --only に渡して python3 .claude/ai-plc-jev/scripts/jev_bt_monitor.py --override-pending accept|reject --layer <Layer> --only X・Y・Z [--except <ID> ...] でまとめて記録する(示した後にできた判定は記録しない=--only を省かない。個別に違う判定は先に --override <ID> reject 等で記録し、--except で除く。記録済みは二重に書かれない)。記録はタスクの実行より先に行う。例外(自動完走中〔RUL_plc_session §10 を指す /goal があるセッション〕): 上の記録(判定ごと・タスクの前)とコピペ用プロンプトへの貼り付けは行わず、完了報告の冒頭に「🧭 未確認の Jev 判定: N件」と貼り付け用1行(python3 .claude/ai-plc-jev/scripts/jev_bt_monitor.py --pending --layer <Layer> の出力から写す)をまとめて出す。人が opt-in した Layer の Jev 判定そのものは自動完走でも実行する。jev_monitor を true にする opt-in(01-collection-jev Phase 6.5)は自動完走中も自動で承認しない(人の回答が無ければ false のまま)。会話監視 hook(.claude/ai-plc-jev/scripts/jev_prompt_hook.py)は 1.12.0-exp.2 から凍結中(新しく登録しない。登録済みなら外してよい)。hook が「💡 Jev会話監視」の行を出しても採用せず、無視してよい。Next Actionのコピペ用プロンプトは /04-operation-jev 形式で書き、Re-Inceptionは /02-inception-jev、Re-Collectionは /01-collection-jev で案内する。

backlog(Agent定義があればそれ)に従い各タスクを実行する最終ステージ。Context Storeからコンテキストを注入し、成果物をArtifact Store(Documents/)に格納する。実行中に発見したコンテキストはContext Storeに追加する(Hierarchical Context Propagation)。

共通規約: 命名は RUL_plc_system §6 / 完了報告は RUL_plc_session §7 / Phase遷移通知は §8 / Mob CP出力は §9 に従う。

自動完走モード: セッション中に RUL_plc_session §10 を指す /goal があるときは §10 に従い、Mob Checkpoint・Next Action・タスク選択で待たずに進める(無ければ本 SKILL のとおり止まる)。

Read the full file on GitHub · 121 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. 16d ago Changed · +5 lines · +8 tokens per session fbc5844b66a8
  2. 18d ago First seen · 116 lines · 53 tokens per session scan A 5122c7a6b9e0

Subscribe to this mod's changes

04-operation-jev is a skill published in the GitHub repository miyatti777/ai-plc (49 stars, last pushed 17d ago), licensed MIT. It adds 61 tokens to every session and 6,072 once invoked, about $0.0002 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 2026-09-30.