Influencer Report

Generate a comprehensive influencer vetting report. Use when someone asks to vet, analyze, or review an influencer/creator. Takes a creator profile URL or vi...

Security Scan

The skill appears to do what it claims (scrape profiles and call Memories.ai APIs) but there are inconsistencies and a surprising external webhook host that warrant caution before installing or supplying credentials.

Purpose & Capability

Name/description, SKILL.md, and the bundled Python scripts align: the code scrapes creator profiles (V1) and submits videos to Memories.ai MAI/metadata endpoints (V2). However, the registry metadata claims no required environment variables while SKILL.md and the code require two Memories API keys (MEMORIES_V1_API_KEY and MEMORIES_API_KEY). That mismatch is unexpected and reduces trust in the packaging/metadata.

Instruction Scope

Runtime instructions are narrowly scoped to scraping a profile and submitting videos to Memories.ai endpoints. They do not instruct reading unrelated local files or other credentials. One noteworthy oddity: poll_result() polls a webhook at https://demo.memories-ai.org/webhooks/memories/result/{task_id} (a different host than the primary API endpoints). Polling an external webhook host for results is unusual and should be verified as legitimate for this API.

Install Mechanism

This is instruction+script-only with no install spec; nothing is downloaded or written automatically by an installer. That keeps install risk low.

Credentials

The SKILL.md and scripts require two API keys (V1 and V2) for Memories.ai — which is consistent with the skill's purpose — but the published registry metadata lists no required environment variables or primary credential. The discrepancy between declared registry requirements and the actual runtime requirements is concerning because a user could install without realizing they'll need to supply API keys. No other unrelated secrets are requested.

Persistence & Privilege

The skill does not request always:true, does not modify other skills, and does not request system-level persistence. It only runs as scripts and makes outbound HTTP requests.

What to consider before installing

Before installing or running this skill: - Do not supply your Memories API keys unless you trust the skill/author. The scripts require MEMORIES_V1_API_KEY and MEMORIES_API_KEY to operate. - Note the registry metadata does NOT list these required env vars even though SKILL.md and the code do — treat that as an indicator the package metadata may be incomplete or sloppy. Ask the publisher to fix the registry metadata. - Verify the endpoints: the code posts to api.memories.ai and mavi-backend.memories.ai, but it polls a webhook at demo.memories-ai.org for results. Confirm with the Memories.ai documentation or the skill author that demo.memories-ai.org is an intended endpoint and not a third-party/attacker host. - Run the scripts in an isolated environment (or sandbox) and inspect network traffic if you must test with real keys. Consider creating limited-scope test keys if Memories.ai supports them. - If you cannot verify the webhook host or the author, avoid installing or give only minimal, revocable credentials and monitor usage.

Influencer Report Skill

Vet influencers by analyzing their recent videos with Memories.ai V1 + V2 APIs.

Setup

Required environment variables:

• MEMORIES_V1_API_KEY — Memories.ai V1 API key (scraper, library, search)
• MEMORIES_API_KEY — Memories.ai V2 API key (MAI transcript, metadata)

Quick Start

From a profile URL (auto-scrapes recent videos):

python3 scripts/influencer_report.py --handle charlidamelio \
  --profile-url "https://www.tiktok.com/@charlidamelio" --scrape-count 5

From direct video URLs:

python3 scripts/influencer_report.py --handle creator_name --platform tiktok \
  --videos https://tiktok.com/@user/video/1 https://tiktok.com/@user/video/2

Workflow

1. Scrape — V1 /scraper ingests creator's recent videos from their profile URL
2. List & Search — V1 /list_videos and /search retrieve the ingested content
3. Analyze — V2 MAI Transcript API provides visual + audio analysis per video
4. Metadata — V2 Metadata API pulls engagement stats (views, likes, comments)
5. Score & Report — Content quality scoring + formatted markdown report

API Endpoints Used

Step	API	Endpoint
Scrape profile	V1	POST /serve/api/v1/scraper
List library	V1	POST /serve/api/v1/list_videos
Search library	V1	POST /serve/api/v1/search
MAI transcript	V2	POST /serve/api/v2/{platform}/video/mai/transcript
Video metadata	V2	POST /serve/api/v2/{platform}/video/metadata

Report Format

See references/report-format.md for the full template and scoring guide.