Agent Acquisition Master -- Agent becomes a master of client acquisition.

WarnAudited by ClawScan on May 10, 2026.

Overview

This sales-outreach skill is openly about acquisition, but it can set up autonomous jobs that send emails, post and interact on social accounts, read Gmail, and persist lead data without clear per-action approval boundaries.

Install only if you intentionally want an automated sales-outreach agent. Before enabling it, use dedicated outreach accounts and domains, review or disable the crons, require previews before sends/posts, define approved lead sources and exclusions, confirm email/social-platform compliance, and decide exactly what may be logged or sent to Telegram.

Findings (5)

Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.

What this means

If enabled, the agent could keep acting on business and social accounts on weekdays after the initial setup.

Why it was flagged

The artifact explicitly describes persistent autonomous scheduled jobs configured during setup, not only user-invoked one-off guidance.

Skill content
7 crons autonomes. Lun-Ven sauf mention contraire. Agent configure ces crons lors du setup du skill.
Recommendation

Require explicit opt-in for each cron, provide a simple disable command, and require user approval before any outgoing public post, email batch, or social action.

What this means

A bad template, target list, or model decision could send unwanted emails, post publicly, or interact with prospects from the user's accounts.

Why it was flagged

The scheduled workflows perform external account actions and scraping/bulk outreach without a documented per-message or per-recipient approval gate.

Skill content
publie le contenu du jour... Twitter + LinkedIn; envoie le batch cold email du jour (max 50); Scrape LinkedIn... identifie 50 nouveaux prospects
Recommendation

Add preview-and-confirm steps, strict recipient/source allowlists, daily caps controlled by the user, and dry-run mode by default.

What this means

The agent may be able to read, send, post, like, connect, or message using the user's live accounts.

Why it was flagged

The skill expects access to logged-in browser sessions for email and social accounts rather than narrowly scoped credentials.

Skill content
virtual-desktop installé (pour Gmail, LinkedIn, Twitter) ... Gmail configuré et session active via noVNC
Recommendation

Use dedicated outreach accounts, narrowly scoped API tokens where possible, and clearly document which accounts the agent may access and what it may do with them.

What this means

Prospect and customer communications could remain in agent memory and influence later tasks beyond the original outreach context.

Why it was flagged

Lead messages, names, context, and campaign results are read from Gmail/logs and persisted into reusable learning memory without retention or exclusion rules.

Skill content
lit les nouvelles réponses Gmail... avec nom, message, et contexte... Lis tous les logs de la semaine... Archive la synthèse dans /workspace/.learnings/LEARNINGS.md
Recommendation

Define what data may be logged, redact personal or sensitive content, set retention limits, and keep .learnings memory scoped to this skill unless the user explicitly shares it.

What this means

The skill may require extra components or accounts that are not obvious from the registry metadata alone.

Why it was flagged

The skill's own metadata names additional skills and optional API keys, while the registry summary says no install spec and no required env vars; users may miss these dependencies.

Skill content
requires: skills: - virtual-desktop - agent-shark-mindset ... env_optional: - HUNTER_API_KEY - APOLLO_API_KEY - BREVO_API_KEY
Recommendation

Declare dependencies and optional credentials in registry metadata and ask users to verify the source and permissions of each dependent skill or provider.