How do I install the Call Scrub skill?

Run npx skillmds@latest add iansteitz1-eng/call-scrub in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Call Scrub safe to use?

This skill has not completed SkillMD's automated safety review yet. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Call Scrub?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Call Scrub free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Call Scrub?

iansteitz1-eng (@iansteitz1-eng) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Call Scrub by iansteitz1-eng — Agent Skill

Name: Call Scrub
Author: iansteitz1-eng

Call Scrub

Pull and analyze your ElevenLabs Conversational-AI voice calls — list recent conversations, or scrub one transcript to surface every tool the agent called, the result it got, errors, and timing. The fast way to debug "why did that call misbehave" without replaying audio. BYOK (your own ElevenLabs API key); read-only; stdlib only. Use when the user says "scrub that call", "what happened on the call", "pull the call transcript", "debug the voice call", "what tools fired".

iansteitz1-eng 4b80458 8 files · 11.1 KB Updated Sep 17, 2026

File contents

call-scrub

Debug a voice call by reading what actually happened — tool calls, results, errors — not by re-listening.

export ELEVENLABS_API_KEY=...                          # BYOK (your key)
python3 call_scrub.py                                  # list recent conversations
python3 call_scrub.py --agent-id agent_xxx --limit 20  # filter to one agent
python3 call_scrub.py --conversation-id conv_xxx       # scrub one call's tool activity

Why

When a Conversational-AI call goes sideways, the audio doesn't tell you which tool returned an error or where it timed out. This reads the transcript via the EL API and lays out every CALL and its -> result, flagging anything that looks like an error/fail/timeout (⚠). We built it debugging our own voice stack; it's the first thing to reach for after a bad call.

Behavior

Default / --list: recent conversations (started, id, msgs, duration, status).
--conversation-id: the full tool-call/result trail + an error count.
No API key → clear FATAL + exit 2. Read-only; nothing is written.

Requires

An ElevenLabs API key (env ELEVENLABS_API_KEY or --api-key). No Python packages.

Frequently asked questions