Call Scorer
Score candidate interview and screening calls against a hiring rubric, using meeting transcripts and notes. Use this skill whenever the user wants to "score the interview", "give feedback on a candidate call", "how did that candidate do", "rate the candidates from this week's interviews", "pull my interview notes and score them", or wants structured feedback on a recorded conversation with a candidate. Connects to Granola via MCP to read transcripts and notes, then scores against the rubric-builder framework. Always score against an explicit rubric — build one first if none exists.
Install
npx skills add getclera/claude-skills --skill call-scorergit clone --depth 1 https://github.com/getclera/claude-skills.git /tmp/claude-skills
mkdir -p ~/.claude/skills
cp -r /tmp/claude-skills/recruitment/skills/call-scorer ~/.claude/skills/call-scorerPersonal skills live in ~/.claude/skills. Use .claude/skills in a project folder to share a skill with your team.
From the SKILL.md
# Interview Call Scorer ## What This Skill Does Reads candidate interview/screening calls (transcripts + notes from **Granola via MCP**) and produces a **structured score and written feedback per candidate**, mapped to the hiring rubric. The output tells the user: how the candidate scored on each dimension, *with evidence quoted or paraphrased from the actual call*, and a clear advance / hold / pass recommendation. This turns a pile of recorded interviews into a ranked, evidence-backed shortlist. --- ## Prerequisites & Limitations (Read First) This skill depends on **Granola MCP**. Know these constraints before relying on it: 1. **Transcript access requires a Granola Business plan or higher.** On the free/Basic tier, the MCP only exposes the last 30 days of notes and *no transcripts.* For real scoring you want transcripts — flag this to the user if scoring seems thin. 2. **Granola MCP only sees the authorizing user's private "My notes" space — never Team-space folders.** If interviews live in a shared Team space, each interviewer must connect their own Granola, or scoring will come back empty. Surface this if a search returns nothing. 3. **A rubric must exist.** Score against an explicit rubric (use the `rubric-builder` skill). Never invent scoring criteria silently — that just launders the scorer's bias. If Granola isn't connected, say so and suggest enabling it. Don't fabricate call content. --- ## Process
Details
- Skill name
call-scorer- Repository
- getclera/claude-skills
- Path
recruitment/skills/call-scorer/SKILL.md- License
- MIT
- Repo updated
- 7 November 2026
FAQ
What does the Call Scorer skill do?
Score candidate interview and screening calls against a hiring rubric, using meeting transcripts and notes. Use this skill whenever the user wants to "score the interview", "give feedback on a candidate call", "how did that candidate do", "rate the candidates from this week's interviews", "pull my interview notes and score them", or wants structured feedback on a recorded conversation with a candidate. Connects to Granola via MCP to read transcripts and notes, then scores against the rubric-builder framework. Always score against an explicit rubric — build one first if none exists.
How do I install the Call Scorer skill?
skills CLI (Claude Code, Cursor, Codex and more): npx skills add getclera/claude-skills --skill call-scorer. Skills go in ~/.claude/skills (personal) or .claude/skills (project), and Claude loads them automatically when a task matches the description.
Is Call Scorer popular?
It lives in getclera/claude-skills, which has 5 GitHub stars and 0 forks.
Related skills
Agentic Todo
5getclera/claude-skills
Set up a personal, persistent todo list inside Claude that fills itself. Claude scans your email, Slack (DMs, mentions, saved messages), meeting notes and calendar on a schedule and drops suggested todos into an inbox you accept or dismiss. Every task has a "Do it" button that hands Claude a ready-made prompt with the context and a plan. Use when the user asks how to track todos, wants a todo list in Claude, wants todos pulled from email/Slack/meetings, or wants an AI task inbox or an agentic todo list.
Application Post Mortem
5getclera/claude-skills
Analyze why a job application or interview didn't work out, and turn it into calibration for the next one. Use this skill whenever someone says "I got rejected", "I didn't get the job", "why am I getting ghosted", "what went wrong with my application", "help me figure out why I keep getting rejected", or wants to learn from a failed process. Reviews the role, the candidate's materials, and what happened — including interview transcripts if they have them (via a note-taker MCP or pasted) — and gives an honest read on the likely cause plus what to change. Gather everything about the process firs
Career Pivot Translator
5getclera/claude-skills
Help someone changing careers translate their existing experience into the language of a new field. Use this skill whenever someone says "I'm switching careers", "how do I move from X to Y", "I want to pivot into [field]", "my experience is in A but I want to do B", "how do I make my background relevant for [role]", or is applying to roles outside their current domain. Finds the genuine transferable value in what they've done and reframes it in the target field's language — honestly, without overclaiming. Gather their full background and the target role before translating.
Chatgpt Agentic Todo
5getclera/claude-skills
Build and maintain a private, persistent agentic task inbox in ChatGPT. Capture actionable suggestions from connected email, Slack, meeting notes, calendar, and recent conversation context; deduplicate and triage tasks; provide a clear actionable UI and a "do it" prompt for each task. Use when the user requests an AI todo inbox, to extract follow-ups from their work, or to recreate Clera's agentic-todo workflow inside ChatGPT.
Bio Read Qc Umi Processing
9hackernoon/cuttlefish-agent-skills
Extracts UMIs and collapses reads to original molecules with umi_tools (directional dedup) or builds error-corrected single-strand/duplex consensus reads with fgbio. Use when the library has UMIs and accurate molecule counting or below-sequencer-floor error correction is needed - single-cell, low-input RNA-seq, targeted panels, and ctDNA/liquid-biopsy rare-variant detection. For UMI extraction during QC use fastp-workflow; do not dedup non-UMI bulk RNA-seq.
Bun Configure Private Registry for Org Scope
9jarle/bun-skills
Configure a private registry for an organization scope with bun install
Browse the most popular Claude skills, all productivity & notes skills, or write your own with the skills generator.