Creator vetting CLI for AI agents

Audit influencers across 5 platforms from the command line. Drop into Claude, Hermes, or any LLM-driven workflow to score creators, detect fake followers, and surface red flags — autonomously.

View on GitHub →

install

# Install the CLI
$ npx skills add collabpal/openclaw
 
# Set your API key
$ export COLLABPAL_API_KEY="cp_live_..."
 
# Audit your first creator
$ openclaw audit @levelsio

Built for automation

Designed for autonomous vetting

A creator-vetting layer for AI agents. Structured output, no SDK, model-agnostic.

AI Agent Ready

Every command outputs structured JSON. No SDK, no parsing — designed for LLMs to plan, execute, and pipe results straight into your workflow.

5 Platforms, 1 CLI

Audit creators on X, Instagram, TikTok, YouTube, and LinkedIn from a single endpoint. We normalize the data so your agent never learns each platform's quirks.

Real Bot Detection

We sample 100 followers per audit and score authenticity — engagement-per-impression, age signals, posting velocity, ratio analysis.

Weighted Sub-Scores

Reach (40%), Audience (30%), Content (20%), Conversion (10%). Composite rounds to 0–100 and maps to a tier S/A/B/C — easy to filter in pipelines.

Red Flags Surfaced

Inflated followers, paid-promo signal, dead engagement, locked accounts — every audit returns a structured array of concerns to act on.

Batch & Stream

Audit a list of handles in parallel, export to CSV or JSON, or stream the results line-by-line into stdout for live agent processing.

Workflow

How the CLI works

From handle to structured verdict in three commands.

Step 01

Discover platforms

List the platforms the CLI can audit and verify your account has access. Each platform has a status (available / coming soon) so your agent can route accordingly.

discovery

$ openclaw platforms:list
 
# JSON output
{
  "platforms": [
    "x", "instagram", "tiktok",
    "youtube", "linkedin"
  ],
  "available": ["x"]
}

Step 02

Audit a creator

Pass a handle or a profile link from any supported platform. The CLI detects the platform, pulls engagement data, samples followers, and returns a structured score.

audit

$ openclaw audit @levelsio
 
{
  "handle": "levelsio",
  "platform": "x",
  "score": 78,
  "tier": "A",
  "subScores": { ... },
  "redFlags": [...],
  "reportUrl": "https://collabpal.com/r/..."
}

Step 03

Batch & integrate

Audit a list of handles, filter by tier, export to CSV/JSON, or pipe the structured output directly into Claude, Hermes, or any LLM-driven workflow.

batch & export

# Batch from a file
$ openclaw audit:batch -f handles.txt
 
# Filter to tier S/A only
$ openclaw audit:list --tier S,A
 
# Export to CSV for a stakeholder
$ openclaw audit:export --format csv > creators.csv

One CLI, every workflow

Audit one creator or a thousand

Same command shape whether your agent runs one audit or batch-scores an entire shortlist.

examples

# Single creator (auto-detects platform from link)
$ openclaw audit "https://instagram.com/levelsio"
 
# Batch from a file, parallel
$ openclaw audit:batch -f handles.txt --concurrency 10
 
# Filter by tier and minimum score
$ openclaw audit:list --tier S,A --min-score 70
 
# Stream JSONL into any agent (Claude, Hermes, ...)
$ openclaw audit:batch -f shortlist.txt --stream | your-agent
 
# Re-audit only stale reports (>24h)
$ openclaw audit:refresh --stale

Reference

Command reference

Eight commands. Same flag shape across all of them.

Command

Description

openclaw audit <handle>
Audit a single creator. Accepts handle (@levelsio) or full profile URL.
openclaw audit:batch -f <file>
Audit a list of creators in parallel. One handle/URL per line.
openclaw audit:list
List past audits. Supports --tier, --min-score, --since, --platform filters.
openclaw audit:get <id>
Fetch a single past audit by its public ID.
openclaw audit:export <id>
Export an audit. Use --format json|csv|markdown.
openclaw audit:refresh
Re-audit stale reports (--stale flag) or specific handles.
openclaw platforms:list
List supported platforms and their availability status.
openclaw config:set <key> <value>
Set API key, default tier filter, output format, concurrency.