ClawHub
Back to skill

Security audit

OpenTweet X Poster

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed OpenTweet integration for managing X posts, media, analytics, and human-reviewed outreach, with guardrails around publishing and DM approvals.

Install only if you trust OpenTweet with the connected X account and API key. Review all post text, schedules, auto-retweet/auto-plug settings, evergreen queue settings, and DM campaign targeting before confirming actions, because approved operations can publish publicly or queue outreach from your account.

A.I.G

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs

SkillSpector

Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • 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)

Natural-Language Policy Violations

Medium
Confidence
94% confidence
Finding
The example ICP for DM campaigns explicitly filters prospects to 'English' speakers without indicating that this is user-provided, optional, or justified by campaign requirements. In an autonomous outreach skill, hard-coded language qualification can cause discriminatory targeting by default and may lead the agent to exclude protected or underserved groups without user awareness.

Autonomous Decision Making

Medium
Category
Excessive Agency
Content
2. `POST /api/v1/dm-campaigns` with the ICP, sources, message brief, limits, and `accept_policy: true`
3. `POST /api/v1/dm-campaigns/{id}/discover`, then poll `GET /api/v1/dm-campaigns/{id}` until status is `reviewing`
4. `GET /api/v1/dm-campaigns/{id}/leads?state=qualified` — show the user the qualified leads and their drafted messages
5. Only after the user approves specific leads: `PATCH .../leads/{leadId}` with `action: approve`. Never auto-approve every lead.
6. `GET /api/v1/dm-campaigns/{id}/inbox` later to read replies

**Find inspiration and repurpose it:**
Confidence
85% 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.

Static analysis

No suspicious patterns detected.