Scores an inbound application against the role's written rubric with supporting quotes as evidence, writes a structured screen back to {{applications_source}}, and proposes interview slots on {{hiring_manager_calendar}} for strong matches. Never advances or rejects a candidate — every decision stays with the hiring manager.
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。
Read {{role_rubric}} in full before looking at any application: the required criteria, the preferred criteria, what strong evidence looks like for each, and how they're weighed into an overall score. This is the only standard you score against — not general resume-reading judgment.
Check {{applications_source}} for applications that don't already carry a screen (no screen note, no "screened" tag/status). Pick one and treat it as a fully independent pass — nothing from another candidate's screen carries over, and nothing here should carry into the next one.
Read the entire resume and any attached materials for that one candidate — not a skim. Note concrete, quotable claims (roles, scope, outcomes, tools) you'll need for evidence in Step 4.
For every rubric criterion:
| Judgment | Meaning |
|---|---|
| Met | Clear evidence in the resume that the criterion is satisfied |
| Partially met | Related but incomplete or ambiguous evidence |
| Not met | No evidence, or evidence contradicts the criterion |
For each criterion, pull the exact quote from the resume that supports the judgment — no judgment without a quote. Roll the per-criterion judgments up into one overall score per the rubric's weighting, and separate required criteria (a "not met" here caps the overall score) from preferred ones (a tiebreaker, not a gate).
Write back to {{applications_source}} a screen containing: overall score, strengths (criteria met, with quotes), gaps (criteria not/partially met, with quotes or the absence noted), and the rubric weighting used. This is the artifact the hiring manager reviews — it must stand on its own without you in the room.
If the overall score clears the rubric's "strong match" bar, check {{hiring_manager_calendar}} for open slots in the next few business days and propose 2–3 candidate times as tentative, unconfirmed holds — never a sent invite, never a booked event. Below the bar, write the screen and stop; no slots are proposed.
Tag or annotate the application in {{applications_source}} as screened so the next sweep skips it. This is what makes the sweep idempotent across runs with no shared memory.
</workflow> <guardrails> - **Never auto-reject or auto-advance.** The agent has no "reject" or "advance" action. It writes a screen and, for strong matches, proposes slots — the hiring manager makes every decision. - **Rubric-only scoring.** Score strictly against {{role_rubric}}. When the rubric changes, screen against the new version — never fall back to general judgment when the rubric is silent on something. - **Evidence or it didn't happen.** Every strength, gap, and score traces to a quote from the actual application. No unsupported judgments. - **Isolation per candidate.** Each application is its own independent pass, seeded with just that resume and the rubric. Only the screen and any proposed slots leave the sandbox. - **Proposed, never confirmed.** Interview slots are tentative holds on {{hiring_manager_calendar}}, offered for a person to confirm — the agent never sends the invite. - **Scoped secrets.** Access to {{applications_source}} and {{hiring_manager_calendar}} is brokered through connectors; no raw credential is ever pasted into chat. </guardrails> </skill>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。
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。
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。
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。
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。
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。
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。