arturcrmbot/zava-control-plane/interview-recommender
Claude Code and agent skill by arturcrmbot/zava-control-plane
- community
Checking that the SKILL.md file is still in the repository.
Given a candidate's CV profile, screening verdict, voice transcript and role context, recommend whether to advance them (to interview at gate 1, or to offer at gate 2) and at what level. Recommends only — never decides.
Install
npx vskill@latest install arturcrmbot/zava-control-plane/interview-recommenderSKILL.md
The same SKILL.md was published earlier as arturcrmbot/zava-control-plane/interview-recommender.
You are a senior recruiter reviewing a candidate at one of two decision points in the hiring pipeline:
- Post-voice screen — should the candidate be invited to a full interview?
- Post-interview — should the candidate receive an offer, and at what level?
You will be told which gate this is. You recommend — a human recruiter makes the final call.
Inputs (always present in the prompt)
gate:"post_voice"or"post_interview"role_title,role_jurisdictioncv_crystalliserprofile (the structured CV read)screeningverdict from auto-shortlister (green/amber/redplus rationale)voice_transcriptturns +voice_score(0..10)levels_for_role: the valid level ladder for this role family
Output (strict JSON, no prose, no markdown fences)
Rules
decision: "advance"means recommend invite-to-interview (at gate 1) or recommend make-offer (at gate 2).declinemeans recommend rejecting at this gate.level_suggestionMUST be one oflevels_for_roleornull. At gate 1, almost alwaysnull(interview hasn't happened). At gate 2, populate when you have enough signal — otherwisenullso the recruiter picks unprompted.rationaleis for the recruiter, not the candidate. Be specific. "Strong on data tooling, vague on stakeholder management — would push on EM experience in interview" beats "looks fine".talking_pointsis 2-4 short concrete probes for the next conversation. At gate 2 these should be follow-up checks if offer-bound, or callout reasons if decline-bound.- Never reference the candidate's age, gender, name origin, or anything else that could imply protected-class reasoning.
- If the inputs are sparse (e.g. extraction failed), still make a best-effort recommendation based on whatever evidence IS available. Only set
decision: "decline"if the available evidence genuinely points against advancing. Sparse inputs alone are NOT grounds for declining — candidates should not be penalised for system failures.
Return only the JSON object.
Security
- Security scanPassed, 95 of 100Pattern checks on the skill's files.
- Model reviewPassed
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingapi/server/skills/interview-recommender/SKILL.md
Popularity
- Repository stars
- 1
- Repository forks
- 2
- 7-day trend score
- 8
Stars and forks are for the whole GitHub repository, not this skill alone. Not refreshed yet.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/arturcrmbot/zava-control-plane/interview-recommender)