「stop」の検索結果

Runs several language models as independent coders on the same labeling or discovery task and reads their disagreement as data — panel assembly for model diversity, keeping votes independent, consensus rules, agreement statistics (Cohen's and Fleiss kappa, Krippendorff's alpha), the correlated-errors caveat that stops agreement being mistaken for validity, human validation beyond the panel, and reporting. Use when the user asks about a council, panel, ensemble, or jury of models, asks how to combine several models' labels, or asks what kappa or alpha to report for model coders. Single-model codebook and validation work goes to text-classification, per-item confidence to llm-calibration-logprobs.

日本語の概要は準備中です。原文の説明を表示しています。

Manage Alibaba Cloud Elastic Compute Service (ECS) using the @alicloud/ecs20140526 TypeScript SDK. Use when working with cloud servers on Alibaba Cloud, including instance lifecycle (create, start, stop, reboot, delete), disks and snapshots, images, security groups, VPC networking, EIP, ENI, SSH key pairs, dedicated hosts, auto provisioning, launch templates, Cloud Assistant commands, tags, system events, diagnostics, storage capacity units, and prefix lists. Covers all 374 APIs of the ECS 20140526 version.

日本語の概要は準備中です。原文の説明を表示しています。

Checks whether each claim in a manuscript is supported by the source it cites, reading the project's per-source Markdown knowledge base. Flags contradicted, unsupported, overclaimed, out-of-scope, and misattributed claims with verbatim source quotes. Runs citation-check first, and stops at a pre-flight gate when the knowledge base is missing, incomplete, or holds unconverted PDFs. Use to verify that a draft's claims match the sources it cites.

日本語の概要は準備中です。原文の説明を表示しています。

Crawl GitHub trending repositories, analyze with LLM for Chinese insights, categorize by themes, compute diffs against history, and generate Markdown reports. Default brief mode stops at trend analysis; optional detailed mode appends per-project analysis. Supports incremental gap-filling and selective re-analysis with caching.

日本語の概要は準備中です。原文の説明を表示しています。

Run the Anabasis v2 improvement loop, the Super Loop: find the one link that holds the climb, choose one change against it, prove its path, put a prediction on record, launch, watch, read the log, decide the next move. Use when the operator asks to keep a campaign going, to run a pass of the loop, what to launch next, why a campaign is not climbing, or what a finished round means for the next one. Owns the choice of experiment and the reading of its evidence; the operator owns promotion, a freeze, a stop, spend and the solve's condition.

日本語の概要は準備中です。原文の説明を表示しています。

Test an uncertain Anabasis v2 path at the smallest real layer that can decide it, before a paid round is asked to: one fact from a log or a dry run, a walk of one value across the layers a change touches, a replay of every recorded log through this tree's readers, a whole round rehearsed on a seeded copy of a campaign with a script in each model seat (`bun run runs seed`, `bun run runs rehearse`), or one live seat from a seed. Use when the operator asks whether a change will work in the real system, to simulate or rehearse a round, to stand again in a position a past run stood in, what a campaign of a given size can read, or what a kill or a stop leaves behind. Not for reading a finished run (whole-run-investigation) or starting a paid one (launch-run).

日本語の概要は準備中です。原文の説明を表示しています。

Stop AI agents from secretly bypassing your rules. Mechanical enforcement with git hooks, secret detection, deployment verification, and import registries. Born from real production incidents: server crashes, token leaks, code rewrites. Works with Claude Code, Clawdbot, Cursor. Install once, enforce forever.

日本語の概要は準備中です。原文の説明を表示しています。

Run and manage ad-hoc AI prompts in sgai workspaces without starting a full agentic session. Use when you need to run a one-off AI prompt against a workspace, check the status of a running ad-hoc prompt, or stop a running ad-hoc prompt.

日本語の概要は準備中です。原文の説明を表示しています。

Inspect and manually purge the Soniox artifacts — transcriptions and uploaded audio files — that offline transcription leaves behind, on dev or prod. Use when offline transcription silently stops working, when Soniox answers 429 `limit_exceeded`, when `SonioxTranscriberTest.OfflineTranscribeWorks` and `CleanerDeletesEnqueuedArtifacts` fail together, or when the user says "sweep Soniox", "soniox/sonoix sweep", "sonoix swipe", "purge Soniox", "check Soniox files", or "is Soniox at the cap".

日本語の概要は準備中です。原文の説明を表示しています。

Use when the user reports a problem or names a piece of work and wants to start on it in a fresh worktree — "start a task", "new task", "file this and make a branch", "/start-task". Finds or files the GitHub issue, creates the feat/… or bugfix/… branch linked to it, and prints the paste-ready `ai fwt|bwt` (Docker) and `ai os fwt|bwt` (host) commands that open the worktree. Linear flow with hard stops; asks nothing.

日本語の概要は準備中です。原文の説明を表示しています。

Run a marketing objective to completion over hours, days, or weeks — not a single deliverable. Takes a business outcome ("40 qualified demos a month from organic by Q4", "cut blended CAC below $90", "launch the product in September"), decomposes it into work items with declared ECO floors, executes them across context windows with resumable state, and stops only when an independent gate returns SHIPPED or CLOSED. Use when the request is an outcome rather than an artifact, when work must continue in the background or across sessions, or when the user says "goal", "run this until", "keep working on", "autonomous", "long-running", "over the next month", or "get us to <number>".

日本語の概要は準備中です。原文の説明を表示しています。

Generate, rank, and gate a reusable hook bank — sourced pains × named hook formula families, scored on clarity/specificity/curiosity/proof-backing, gated at the 10/16 ad threshold, delivered in 3 lengths with per-hook provenance and downstream routing. Use when "hook bank", "generate hooks", "hook generator", "write hooks for", "ad hooks", "scroll stoppers", "opening lines", "hook ideas", "first 3 seconds", "hook matrix", "I need 50 hooks", or any request to batch-produce attention-openers for ads, social, or video.

日本語の概要は準備中です。原文の説明を表示しています。

当用户在制定战略、做决策、或经营企业/团队时，需要一套"不该做什么"的排除框架。 典型触发：用户说"我们什么都想试""机会很多但不知道怎么选""公司发展太快有点失控""怎么避免致命错误"。 不适用于：纯粹的任务规划（要做清单）、信息查询类问题。 核心逻辑：不做什么比做什么更重要——通过主动建立"不为清单"来防止致命错误。

日本語の概要は準備中です。原文の説明を表示しています。

当投资者因股价涨跌产生焦虑、恐惧或贪婪情绪，需要回归理性决策时激活。 典型触发：股价暴跌时恐慌想卖、暴涨时焦虑要不要止盈、与他人比较收益而心态失衡、因用了非闲钱而睡不着觉。 语言信号："我该不该止盈""跌了这么多我好慌""别人赚了那么多我没赚到""心跳跟着股价走"。 不适用于：纯信息查询（问某公司财报数据）、明确的买入/卖出决策逻辑（那是能力圈或机会成本skill）、短期交易策略咨询。 与相邻skill区分：能力圈skill解决"懂不懂"的认知问题，本skill解决"懂了但情绪仍然失控"的情绪-认知断裂问题；Stop Doing List是事前纪律防线，本skill是事中情绪恢复机制。

日本語の概要は準備中です。原文の説明を表示しています。

Write or rewrite prose so it sounds like a human wrote it — not a frontier model. Named in homage to Rabindranath Tagore, whose prose carried what frontier models reach for and miss: a point of view, specificity over abstraction, and restraint over puffery. Merges two complementary approaches: a 29-pattern catalog of AI tells (from humanizer) plus an 8-rule operating system with an 8-dimension scoring gate (extending stop-slop). Use when drafting, editing, or reviewing any prose: essays, posts, docs, reports, emails. Detects and removes inflated symbolism, promotional language, superficial -ing analyses, vague attributions, em dash overuse, rule of three, AI vocabulary, passive voice, negative parallelisms, filler phrases, inanimate-verb constructions, narrator-from-a-distance voice, and metronomic rhythm. Adds back the things AI writing usually lacks: point of view, stakes, specificity, restraint, varied rhythm, and trust in the reader.

日本語の概要は準備中です。原文の説明を表示しています。

Use when running ANY terminal command - enforces clear progress logging, output reading, and error-stop behavior so terminal processes are never left unchecked

日本語の概要は準備中です。原文の説明を表示しています。

Active Directory exploitation after domain credentials exist: Kerberos (AS-REP/Kerberoast including cracking the ticket then using the account), delegation, NTLM coercion/relay, lateral movement, ACL abuse, ADCS ESC1-ESC17 and CVE paths, dMSA/BadSuccessor, Kerberos reflection, identity confusion, management-plane, domain trust, and domain persistence. Own the current AD chain through crack-and-use to DA, equivalent domain control, or the targeted host SYSTEM. Do not stop after requesting TGS. Host C2 belongs to /post. Operator chooses next modules.

日本語の概要は準備中です。原文の説明を表示しています。

Client-side initial access and social-engineering attack module. Use when the operator selects a human/client/browser/identity-delivery attack direction: ClickFix/FileFix, AiTM session phishing, OAuth device-code or consent abuse, Teams/helpdesk social engineering, HTML/file delivery, QR/mobile handoff, legacy HTA/OLE/Office paths. This is an Attack module: own the chain until host shell or remote interactive foothold. Credential/token/session are intermediate results, not the stop condition. If endpoint AV/EDR/AMSI/application-control blocks an otherwise-valid client execution path, the operator may temporarily select /edr-bypass and then resume this phishing chain until shell.

日本語の概要は準備中です。原文の説明を表示しています。

PencilPlaybook is the UI Skills / Taste-Skill for Pencil.dev + Claude Code — a design playbook that gives Claude real perceptual psychology and senior-level guardrails so it stops producing averaged-out AI slop.

日本語の概要は準備中です。原文の説明を表示しています。

Use when real user interviews aren't possible (solo/hobby/dogfood projects) but persona work is still needed. Enforces epistemic discipline: adversarial spectrum, pre-committed stop conditions, speculation tagging. NOT a substitute for /mycelium:user-interview when real users are available.

日本語の概要は準備中です。原文の説明を表示しています。

Reference for Claude Code's 6 Task Management tools — TaskCreate, TaskGet, TaskList, TaskUpdate (structured tracking) and TaskOutput, TaskStop (background execution). Covers tool parameters, status lifecycle, completion rules, dependency management, and conventions

日本語の概要は準備中です。原文の説明を表示しています。

Manage Claude Code --channels sessions (start/stop/status/health) via channel-manager.cjs and terminal-tracker.cjs.

日本語の概要は準備中です。原文の説明を表示しています。

This skill should be used when the user wants to build, set up, or operate an AI "second brain" — a persistent memory vault (Obsidian-compatible Markdown) that lets an AI agent resume work without re-explaining context, remembers correction instructions across sessions, and auto-organizes raw material into structured knowledge. Trigger on requests like "build my second brain", "set up an Obsidian vault for AI", "give my AI persistent memory", or "make AI stop forgetting".

日本語の概要は準備中です。原文の説明を表示しています。

Strip AI writing patterns from prose across five registers — casual, professional, academic, technical, and narrative. Applies 200+ pattern rules to eliminate predictable LLM tells. Triggers on clean up writing, remove AI slop, edit prose, make this sound less AI.

日本語の概要は準備中です。原文の説明を表示しています。