ClawHub
Back to skill

Security audit

Social Media Scheduler

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed social media scheduling integration that can publish through user-authorized accounts, with explicit approval steps before live posting.

Install only if you are comfortable connecting Socialync and letting the agent prepare, schedule, or publish posts to social accounts you have already authorized. Review each draft carefully before approval, and use the recurring autonomous mode only for content streams where unattended posting is acceptable.

SkillSpector

Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (2)

Autonomous Decision Making

Medium
Category
Excessive Agency
Content
1. `list_profiles`: a user may manage several brands. Never assume the default.
2. `check_quota`: returns the plan, remaining posts, and per-platform daily
   limits. Publishing without checking is how you get a partial batch.
3. `list_connections`: confirm the target platform is actually connected and
   healthy before you draft against it.
4. `create_post_draft` then `schedule_post_draft`, or `publish_now` for immediate
Confidence
75% confidence
Finding
Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.

Behavior Manipulation

Medium
Category
Prompt Injection
Content
and the connection can be revoked from Socialync account settings at any time.

Publishing works on the free plan: $0, no credit card, 5 posts a month across all
8 platforms. Never tell the user they need to pay before they can connect.

Posts go out only through social accounts the user has already authorized inside
Socialync. The agent cannot reach an account the user has not connected.
Confidence
75% confidence
Finding
Subtle instructions detected that may alter agent decision-making or introduce hidden biases.

Static analysis

No suspicious patterns detected.