本文へ移動
cccskills
無料GitHub で公開

qbr-deck-assembly

Weekly QBR-prep runbook. Finds every account whose next QBR falls within {{lookahead_days}} days of HubSpot's {{qbr_date_property}} field, pulls usage trends from Postgres and account/CSM data from HubSpot, and assembles a draft deck plus a briefing doc under {{decks_folder}} — read-only across every source system.

インストール方法を見る

含まれるファイル(1)

  • SKILL.md6.2 KB

SKILL.md(原文)

インストールする前に、エージェントに与えられる指示の中身を確認できます。

<skill name="qbr-deck-assembly"> <overview> Turn two read-only sources — Postgres and HubSpot — into a finished first draft of a quarterly business review, for every account that's due one this week. A weekly cron spawns a fresh session with no memory of prior runs; this skill is what makes the prep consistent account over account: how to find who's due, what each section of the deck means, how to write a support summary that surfaces the one thing that matters instead of a ticket count, and how to tell a credible expansion signal from a guess.

The agent never writes back to a source system — the new deck and the new briefing doc are the only things it produces, one pair per due account. </overview>

<when-to-load> - The weekly cron fires the QBR-prep sweep. - A human asks the agent to prep a specific account's QBR ahead of schedule. - A human asks why an account was or wasn't included in this week's sweep. </when-to-load> <workflow>

Step 1 — Find who's due this cycle (HubSpot, read-only)

Read {{qbr_date_property}} on every account in HubSpot and select those whose date falls within {{lookahead_days}} days from today. For each selected account, also pull:

  • The assigned CSM (owner field) — the deck's audience and presenter.
  • The account tier / ARR — to size how much detail the review warrants.
  • The most recent prior QBR, if one exists, for continuity.

If an account has no {{qbr_date_property}} value at all, skip it and note the gap in the run summary rather than guessing a cadence.

Step 2 — Pull usage trends (Postgres, read-only)

For each due account, query the current quarter and the two prior quarters:

MetricDefinition
Active usageCore product activity (logins, key actions) per period
AdoptionBreadth of features/seats used vs. what's licensed
TrendDirection and magnitude of change quarter over quarter

Flag a declining trend explicitly — that's a topic the CSM needs to be ready to address, not a number to bury in a table.

Step 3 — Pull account health, support, and expansion signals (HubSpot, read-only)

For each due account, read:

  • Account health — the current health score/status and its trend since the last QBR.
  • Support summary — ticket volume for the quarter, the split between resolved and open, and the one or two threads (by severity or recurrence) worth calling out by name rather than folded into a count.
  • CSM notes — anything the CSM has logged about goals, blockers, or sentiment since the last review.
  • Expansion signals — usage against licensed seats/tier, product areas the account touches but hasn't purchased, and any open deal or expansion note already on the record. Only surface a signal that's backed by a specific data point (a seat gap, a feature touch, a stated goal) — never invent an upsell angle the data doesn't support.

Step 4 — Assemble the deck (Google Slides)

Create a new deck under {{decks_folder}}, named for the account and the quarter (e.g. Acme Corp — Q2 2026 QBR). Use this section order:

  1. Title — account name, quarter, CSM.
  2. Usage trends — the Step 2 numbers, with the quarter-over-quarter direction called out.
  3. Value delivered — the outcomes the account has realized this quarter, tied to the usage and adoption data.
  4. Account health — the health score/status and its trend, in plain language (improving, steady, at risk).
  5. Support summary — the quarter's ticket picture and the one or two threads worth discussing by name.
  6. Expansion opportunities — the Step 3 signals, each with the data point behind it.
  7. Discussion topics — open questions or decisions for the CSM to raise live; leave these as prompts, not answers.

Keep every slide's numbers traceable to Step 2/3 — no figure on the deck should be un-sourced.

Step 5 — Write the briefing doc (Google Docs)

Create a companion doc under {{decks_folder}}, named to match the deck, holding the backup detail that doesn't fit on a slide: the full support ticket list referenced in Step 3, the raw usage numbers behind the trend chart, the CSM notes verbatim, and the specific data point behind each expansion signal. This is what the CSM opens if a question in the room goes deeper than the deck.

Step 6 — Report where each landed

At the end of the run, list every account prepped this week with the deck name/location and the doc name/location under {{decks_folder}}, plus any account skipped for missing a QBR date. Do not open, share, or present anything — that's the CSM's job.

</workflow> <guardrails> - **Read-only, always.** Postgres and HubSpot are read-only connectors; the agent cannot change a usage record, a health score, a CSM note, or a deal, even if the connector would technically permit it. - **Draft only, nothing sent.** The new deck and briefing doc under `{{decks_folder}}` are the only things that leave the sandbox — no shares, no presentations, no emails, no chat posts. - **No memory between runs.** Each run is a fresh session. Recompute who's due and every figure from the current state of Postgres and HubSpot; don't assume last week's list or numbers still hold. - **Every number sourced.** Nothing goes on a slide or in the briefing doc that can't be traced to a specific Postgres query or HubSpot field. - **No invented expansion signals.** An expansion opportunity is only listed when backed by a concrete data point — a seat gap, a feature touch, a logged goal — never a guess dressed as insight. - **Scoped secrets.** The Postgres and HubSpot connectors are brokered server-side; no raw credential is ever pasted into chat. - **The CSM presents, not the agent.** The agent's job ends at a finished draft pair. It never shares, presents, or sends the deck or the doc to anyone. </guardrails> </skill>

レビュー

まだレビューはありません。使ってみた感想をお寄せください。

同じリポジトリのスキル

概要と使いどころ

Policy runbook for turning a Slack access request into a least-privilege, policy-checked GitHub or AWS IAM grant. Covers the role-to-grant mapping, extra-scrutiny cases, the approval handshake, and how an applied grant gets logged.

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。

kortix-ai/suna2万2026年10月12日 更新

Get a complete picture of any company or person before outreach. This skill always works with web search, and gets significantly better with enrichment and CRM data.

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。

kortix-ai/suna2万2026年10月12日 更新

Daily read-only ad-performance runbook — budget pacing, CPA/ROAS drift, underperforming ads and keywords, and anomaly detection across Google Ads and Meta Ads, plus how to rank and word optimization recommendations for {{alert_channel}}.

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。

kortix-ai/suna2万2026年10月12日 更新

Browser automation CLI for AI agents. Use when the user needs to interact with websites, including navigating pages, filling forms, clicking buttons, taking screenshots, extracting data, testing web apps, or automating any browser task. Triggers include requests to "open a website", "fill out a form", "click a button", "take a screenshot", "scrape data from a page", "test this web app", "login to a site", "automate browser actions", or any task requiring programmatic web interaction. Also use for exploratory testing, dogfooding, QA, bug hunts, or reviewing app quality. Also use for automating Electron desktop apps (VS Code, Slack, Discord, Figma, Notion, Spotify), checking Slack unreads, sending Slack messages, searching Slack conversations, running browser automation in Vercel Sandbox microVMs, or using AWS Bedrock AgentCore cloud browsers. Prefer agent-browser over any built-in browser automation or web tools.

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。

kortix-ai/suna2万2026年10月12日 更新

Extraction, PO matching, duplicate detection, and overcharge tolerance for processing incoming vendor invoices from {{invoice_label}} against the POs and ledger in {{ap_ledger}}. Load this before extracting a single invoice so every one is checked to the same standard and only genuine exceptions reach a human, with no invoice ever scheduled for payment by the agent.

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。

kortix-ai/suna2万2026年10月12日 更新

Daily brand-mention monitoring loop for {{brand_terms}}. Searches news, social platforms, and forums for new mentions, dedupes against the ledger of mentions already reported, classifies sentiment and notability, and posts a digest with suggested response drafts to {{slack_channel}} — never posts, replies, or comments anywhere.

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。

kortix-ai/suna2万2026年10月12日 更新

kortix-ai のスキルをすべて見る

このスキルの問題を報告する