ClawHub
Back to skill

Security audit

Word Document Composer

Security checks for vulnerabilities and agentic risk

Overview

The skill’s document features match its description, with disclosed local storage and no network calls in either document action.

Inspection reports contain extracted document text, so protect or delete exports when handling confidential documents. The .mochilab/docx-composer directory retains a local database. Keep the host sandbox enabled, visually check generated documents, and review connection permissions separately if choosing the external assistant installer.

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
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Findings (6)

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
90% confidence
Finding

The supplied code is shared infrastructure rather than a Word document composer. Workspace validation, persistence, and reporting could support the declared purpose and are not independently treated as mismatches. However, the general-purpose network request helper is not explained by the description, and the chunk does not implement any DOCX functionality. The network helper is only defined here; there is no evidence that it is invoked or that data is transmitted. Other, unsupplied files could implement the declared document features.

Content

No source excerpt is available for this finding.

Skill declares no tool scope ('permissions' or 'allowed-tools') but code capabilities were detected: file_read, file_write, network, shell.

Medium
Category
MCP Least Privilege
Confidence
70% confidence
Finding

Without declared permissions the skill's intent is opaque and cannot be validated.

Content

No source excerpt is available for this finding.

Workspace state is created and persisted without disclosing the local storage side effect.

Medium
Category
Output Handling
Confidence
90% confidence
Finding

Constructing Store silently creates a .mochilab directory and SQLite database, and put persists supplied records there. There is no confirmation, user-facing log, or explanatory comment/docstring disclosing this storage; runmain constructs Store for every request.

Content

No source excerpt is available for this finding.

The DOCX composer includes a general-purpose outbound HTTP client unrelated to document creation or inspection.

Medium
Category
Excessive Agency
Confidence
97% confidence
Finding

The manifest describes creating DOCX documents and inspecting their text structure, but safe_request implements outbound network requests, including POST, PUT, PATCH, and DELETE with caller-supplied headers and bodies. This capability is not justified by the stated document-processing purpose; the supplied file defines it but does not show it being invoked.

Content

No source excerpt is available for this finding.

HTTP requests transmit caller-supplied data without a disclosure of the transmission.

Medium
Category
Output Handling
Confidence
84% confidence
Finding

safe_request sends supplied headers and body to the selected network endpoint without a confirmation, user-facing log, or explanatory comment/docstring describing that transmission. HTTPS and destination validation protect the connection but do not disclose the outbound data transfer.

Content

No source excerpt is available for this finding.

Scope Creep

Low
Category
Excessive Agency
Confidence
70% confidence
Finding

Skill's behavior or capabilities extend beyond its stated purpose. Scope creep allows an agent to perform actions unrelated to its documented functionality, increasing the attack surface.

Content

Scanner excerpt · LICENSE (reported line 7)May include surrounding context.

text
Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.

Static analysis

No suspicious patterns detected.