「human review」の検索結果

Produces a draft controls and gap pack from a requirement, regulation, standard or policy and the organisation's control descriptions: obligations broken out of the source text, the controls that appear to address each, apparent coverage, gaps, questions for control owners and actions to assess. Never concludes compliance or that a control is effective; owners and audit assess. Use when the user asks to "map this regulation to our controls", "run a controls gap analysis", "break this standard into obligations" or "where is our control coverage thin". Do not use for comparing a policy document against a standard, use policy-gap-review instead; to draft evidence requests from the mapped controls, use control-evidence-request-pack. Drafts for human review; never approves, authorises or signs off.

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。

Drafts answers to the questions in a received RFP, RFQ or tender from the organisation's past responses, answer library, case studies, policies and other documents provided: one row per buyer question with a draft answer marked Reused, Adapted or Missing, its source and date, and a legal or pricing review flag on every answer touching terms, liability, warranties, insurance, data protection or price. Use when the user asks to "draft our RFP response", "answer this tender from our past proposals", "pre-fill the bid questions", "first pass at the RFQ answers" or "reuse our previous bid content". Do not use for writing the buyer's own RFP or requirements, use rfp-requirements-pack instead; for security or due diligence questionnaires use vendor-security-questionnaire-prefill; for scoring received supplier responses use rfp-comparison-pack. Drafts for human review; never approves, authorises or signs off.

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。

Turns a schedule extract (milestone list or planning tool export) and the period's status notes into a DRAFT plain-language slip explanation: which milestones and activities slipped and by how many working days, arithmetic shown, why per the notes as verbatim fragments with source and date, what each slip pushes next per the extract, and the decision each slip needs framed as a question with the options the notes propose and the owner the governance documents name, every missing fact marked UNKNOWN. Use when the user asks to "explain the schedule slip", "what slipped this month and why", "write the schedule commentary for the steering pack", "put the delay in plain language for the sponsor" or "what decisions do these slips need". Do not use for the weekly status report, use project-status-tracker instead; for cost variances, use budget-variance-explainer; for one sprint's carry-over, use sprint-review-summary. Drafts for human review; never approves, authorises or signs off.

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。

Prepares a discovery call brief from what the user has on a prospect (CRM record, inbound form, emails, notes, public material the user supplies): what is known with sources, gaps against the qualification dimensions, hypotheses to test, questions in priority order and what to listen for. Returns a DRAFT Markdown brief in the chat. Use when the user asks to "prep me for the discovery call with <prospect>", "what should I ask on this first call", "intro call brief", "questions for the qualification call" or "what do we know before we speak to them". Do not use for scoring leads against a rubric, use lead-qualification-scorer instead; for an existing customer's account review use account-plan-builder. Drafts for human review; never approves, authorises or signs off.

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。

Prepares a DRAFT vendor risk screening brief from the material the user provides on one vendor (questionnaire answers, certificates, assurance reports, policies, contract extracts): per screening topic, what a document evidences with reference, date and scope, what is only asserted, what is missing or contradicted, certificate and report details as stated, and ready-to-send questions for the vendor and the internal owners. No web research, no risk rating, tier, score or recommendation. Use when the user asks to "screen this vendor", "review the vendor's questionnaire answers", "check what the vendor's certificate covers", "what is missing from this due diligence pack" or "draft the follow-up questions for the vendor". Do not use for answering a customer's questionnaire about your own organisation, use vendor-security-questionnaire-prefill instead. Drafts for human review; never approves, authorises or signs off.

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。

Applies a change-safety review to any requested file, mail or calendar change: lists every create, modify, move, rename, archive or delete before proposing it, marks destructive items as requiring the user's typed confirmation, proposes versioned file names instead of overwrites, declines to draft bulk destructive operations, and drafts a change journal entry for the user to keep. Use when the user asks to "clean up this folder", "delete the old versions", "reorganise these files", "archive everything older than a year", "rename and move these documents" or "make sure nothing gets overwritten while we work", or when any task is about to create or modify files, emails or calendar items. Do not use for backing up, journalling or restoring a skills folder, use skills-backup-keeper instead. Drafts for human review; never approves, authorises or signs off.

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。

Turns the text of a regulatory change (new regulation, amendment, rule, guidance or standard revision) into a DRAFT impact note: what changed against the prior text where supplied, who is affected as the text scopes it, an obligations table with dates and owners to confirm, and open interpretation questions for legal. States only what the text says; the rest is UNKNOWN. Use when the user asks to "brief legal on this new regulation", "list what we must do by when under this amendment", "compare the new standard with the version it replaces", "turn this directive into an obligations table" or "what does this rule mean for us". Do not use for assessing existing controls against a requirement set, use controls-gap-pack instead; for gapping a policy against a standard, use policy-gap-review. Drafts for human review; never approves, authorises or signs off.

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。

Drafts one knowledge article or known-error article from a resolved incident or ticket thread: symptoms as reported, environment, cause exactly as the resolver stated it, resolution steps as performed, workaround with its limits, verification, things tried without effect, related records, keywords and proposed metadata, every statement traced to a ticket note or marked UNKNOWN, with a redaction log and a question list for the knowledge owner. Never decides a root cause, adds a step or publishes. Use when the user asks to "write a KB article from this ticket", "turn this resolved incident into a knowledge article", "draft a known error record for this", "document the fix we just did" or "make this thread into a how-to for the service desk". Do not use for a procedure the team will execute (runbook, operating procedure, playbook), use runbook-drafter instead; for a post-incident review, use incident-postmortem-drafter. Drafts for human review; never approves, authorises or signs off.

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。

Turns an engineer's notes, a ticket history or a chat thread into a draft step-by-step runbook with preconditions, stop conditions, numbered steps with checks, verification, rollback and escalation, for the team to validate before use. Commands verbatim, placeholders marked, every gap UNKNOWN; never executes a step or fills a gap from general knowledge. Use when the user asks to "write a runbook from these notes", "turn this resolved ticket into a runbook", "document the steps we ran" or "check this runbook for gaps". Do not use for a business process without commands, use sop-drafter instead; for a knowledge article or known-error record, use knowledge-article-drafter; for the post-incident review, use incident-postmortem-drafter. Drafts for human review; never approves, authorises or signs off.

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。

Turns one decision discussion (meeting transcript, chat thread, email chain or notes) into a DRAFT one-page decision memo: the decision as stated, status, options considered, evidence cited, dissent recorded verbatim, owner, effective date and review date, with UNKNOWN for anything the discussion did not settle. Use when the user asks to "write up the decision", "record what we decided", "capture the decision from this thread", "draft the decision memo" or "add this to the decision log". Do not use for full minutes or action items, use transcript-to-actions instead; for a technical design decision, use architecture-decision-record; for a paper seeking a decision not yet taken, use board-paper-skeleton. Drafts for human review; never approves, authorises or signs off.

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。

Flags where a P&ID draft raises a process safety question and points to the candidate evidence an engineer would review (HAZOP register, relief study, SIL assessment, isolation philosophy), as questions and quoted passages with UNKNOWN for missing documents, plus safety functions, relief devices and isolation requirements carried over as stated. Never rates, sizes, classifies, resolves or declares protection adequate. Use when the user asks to "write the process safety flags", "produce the process safety section of the P&ID enrichment", "list which protection questions this P&ID raises" or "map the HAZOP and relief study evidence to the draft" after gate G2, once the piping and instrumentation sections exist. Do not use for proposing loops, alarms or trips from the control philosophy, use instrumentation-and-control-enrichment instead; do not use for checking identifiers, use tagging-and-numbering. Drafts for human review; never approves, authorises or signs off.

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。

Assembles a DRAFT capital expenditure request pack from the requester's inputs: business need, options including do nothing, costs exactly as provided, benefits as stated, risks, assumptions, budget status and the approvals route read from the organisation's own delegation of authority, with UNKNOWN for every missing figure, date or name. Use when the user asks to "prepare a capex request", "draft the business case for this equipment", "fill in the capital request form", "structure this investment proposal" or "what approvals does this spend need". Do not use for a board or committee decision paper, use board-paper-skeleton instead; for an IT change approval, use change-request-pack; to pressure-test a finished business case, use cfo-reviewer. Drafts for human review; never approves, authorises or signs off.

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。

Reads one spreadsheet, CSV export or pasted table together with the user's question and drafts an insight pack: a data profile, observations cited to named columns and values and framed as leads, caveats, and suggested charts and follow-up analyses. Every figure carries the label "as read, verify in source"; correlation is never stated as causation. Use when the user asks to "profile this spreadsheet", "what does this data say", "find anomalies in this extract", "give me a first read on this CSV" or "summarise this table". Do not use for a purchase order log or spend extract, use purchase-order-anomaly-review instead; for actuals versus budget, use budget-variance-explainer instead; for a report that arrives by email on a schedule, use report-attachment-analyzer instead. Drafts for human review; never approves, authorises or signs off.

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。

Proposes redactions of personal identifiers in pasted text or an attached document (names, contact details, identification and account numbers, dates that identify, quasi-identifiers) and returns a DRAFT redacted copy with consistent placeholder tokens plus a redaction table stating what was replaced, where and why, with a confidence level and a blank reviewer decision per row. Use when the user asks to "redact the names in this document", "de-identify this transcript before we share it", "strip personal data from these notes", "anonymise this report for publication" or "what would need redacting here". Do not use for assessing a planned processing activity, use dpia-draft-pack instead; do not use for working out what data an incident exposed, use data-incident-impact-brief instead. Never declares a document anonymised, safe to release or compliant. Drafts for human review; never approves, authorises or signs off.

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。

Turns one long support ticket thread (helpdesk history, email chain, chat transcript, call notes) into a one-page DRAFT escalation summary: the customer's ask in their own words, account context as stated, a dated history, what was tried and the stated result of each attempt, the current state, and what is needed from whom by when, every line referenced to a message and every missing fact marked UNKNOWN. Use when the user asks to "escalate this ticket", "summarise this thread for tier two", "write the hand-over for this case", "what has been tried so far", "brief engineering on this customer issue" or "prepare the escalation note". Do not use for a batch of tickets, use ticket-triage-pack instead; for the review after an outage, use incident-postmortem-drafter. Drafts for human review; never approves, authorises or signs off.

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。

Triages a batch of customer support tickets (helpdesk export, shared mailbox, chat transcripts, web forms) into one row per ticket with category, urgency graded on the evidence in the ticket, suggested next action and owner, plus a DRAFT customer reply for every ticket that needs one, returned as one Markdown pack. Use when the user asks to "triage these tickets", "sort the support queue", "what needs an answer first", "categorise this backlog", "draft replies for these tickets" or "build today's queue review". Do not use for internal requests from colleagues, use request-intake-triage instead; for one long thread that needs handing over, use escalation-summary; for a personal mailbox, use inbox-triage. Drafts for human review; never approves, authorises or signs off.

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。

Find bugs, security vulnerabilities, and code quality issues in local branch changes. Use when asked to review changes, find bugs, security review, or audit code on the current branch.

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。

Get a second opinion from an external LLM at three checkpoints: T1 is the plan valid, T2 which path at a fork, T3 finish plus a QA breaker that tries to break the result. One structured move per exchange (PROPOSE, COUNTER, VERIFY, ACCEPT, BLOCK) with memory across turns, and every exchange mirrored into a human-visible channel. Supports a multi-vendor panel instead of a single reviewer. Triggers: "/secondop", "get a second opinion", "run the review panel".

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。

実装計画（plan.md / pbi-input.md）に含まれる危険操作・触れてはならないスコープ・人間承認必須条件を検出し、AI 自律実行前のゲートとして機能する

Use OpenAI Codex CLI as a peer collaboration partner — same standing as Claude, not a one-shot tool. Bring Codex in EARLY (ideation, framing decisions, stuck moments), not only for late peer-review. Hand it strategic context (target, research, draft, reasoning), iterate conversationally over ~3-8 turns, converge before surfacing to the human. Invoke when the user says "check / iterate / ideate with Codex" or "ask Codex" — and proactively when stuck between two reasonable framings, pressure-testing a load-bearing claim or structure, or about to ask the human to pick between options two agents could resolve.

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 just says "plan", "planen" or "roadmap": offer a template via aos-plan-canvas templates. Open plans and HTML artifacts in a local browser canvas where the human annotates elements, chats, and approves or requests changes without leaving the page. Use for planning, whenever a user says "plan", "roadmap", "visualize the plan", "review the plan" or /plan-canvas, right after a brainstorm or grill interview has produced a plan, and whenever any skill's artifact (plan, design, comparison, report, diagram) can be reviewed visually instead of with typed feedback.

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。

Discuss a community-edition product requirement with the user, draft a bilingual (Chinese + English) PRD for review, then — after explicit user approval — create the English-only issue on GitHub (<OWNER/REPO>). Use whenever the user wants to turn a requirement (from a Feishu doc, a meeting note, a customer ask, or a rough idea) into a GitHub issue. Triggers on phrases like "create requirement doc", "create a requirement", "put this requirement on GitHub", "create new feature issue", "draft an issue for...", or anytime the user describes a product feature/improvement they want tracked. The format is opinionated (Background / Goal / Spec / Verification / Related, NO implementation details, draft shown bilingually for human review, issue itself written in English only).

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。

Multi-agent autonomous startup system for Claude Code. Triggers on "Loki Mode". Orchestrates 100+ specialized agents across engineering, QA, DevOps, security, data/ML, business operations, marketing, HR, and customer success. Takes PRD to fully deployed, revenue-generating product with zero human intervention. Features Task tool for subagent dispatch, parallel code review with 3 specialized reviewers, severity-based issue triage, distributed task queue with dead letter handling, automatic deployment to cloud providers, A/B testing, customer feedback loops, incident response, circuit breakers, and self-healing. Handles rate limits via distributed state checkpoints and auto-resume with exponential backoff. Requires --dangerously-skip-permissions flag.

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 Busabase-backed sales prospecting and first-touch outreach desk with a bundled Hono App-in-Skill deployable to AirApp. Use whenever the user invokes $kelly-sales-outreach or /kelly-sales-outreach, wants to enter a product or service and discover its ideal customers, asks to continuously find target accounts or public business contacts, needs evidence-based ICP scoring and personalized outreach drafts, or wants to review and send cold outreach safely. Four subcommands - setup, profile, research, send - plus the desk itself. Every first-touch message requires human approval; one company receives one initial email; guessed addresses, purchased lists, secret exposure, automated spam, and unreviewed external sends are forbidden.

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。