ClawHub
Back to skill

Security audit

BulkPublish: Manage Channels

Security checks for vulnerabilities and agentic risk

Overview

This Markdown-only skill is not malicious, but it documents organization channel-set create, update, and delete actions beyond its stated channel-listing and health-check purpose.

Install only if you want the agent to help manage BulkPublish channel sets, not just inspect connected accounts. Before allowing any create, update, or delete action, verify the exact organization, channel-set ID, channel list, and whether the change is reversible in BulkPublish.

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
  • Tool MisuseTool Parameter Abuse, Chaining Abuse, Unsafe Defaults
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (2)

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
88% confidence
Finding

The skill includes destructive channel-set deletion and other mutable REST operations despite being framed as a channel-inspection skill. Providing direct delete semantics in the skill context can lead an agent to perform parameterized state-changing actions on organization resources if prompted ambiguously or maliciously, causing loss of saved targeting configurations.

Content

Scanner excerpt · SKILL.md (reported line 42)May include surrounding context.

md
| `GET /api/channel-sets` | List sets (ordered by name) |
| `POST /api/channel-sets` | Create — body `{name, channelIds}` |
| `PUT /api/channel-sets/{id}` | Rename or change channels — body `{name?, channelIds?}` (send at least one) |
| `DELETE /api/channel-sets/{id}` | Delete → `{success: true}` |

- Set object: `id`, `name`, `channelIds` (number[]), `createdAt`, `updatedAt`
- `name` is trimmed, max 100 chars, unique per organization — a duplicate name returns **409** with `error.code: "DUPLICATE_NAME"`

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The skill is described as only listing connected channels and checking health, but the documentation also exposes additional capabilities including mention search and full channel-set management. This scope expansion increases the chance an agent will invoke unintended tools or perform state-changing actions outside the user's expected intent, violating least privilege and enabling unauthorized account targeting changes.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.