ClawHub
Back to skill

Security audit

File Diff and Delivery Review

Security checks for vulnerabilities and agentic risk

Overview

The skill performs local file comparisons and delivery checks, with disclosed local storage and no active network or credential-access workflow.

Choose only the files or folders needed for comparison. Reports can include private source content, so review them before sharing. Local outputs remain under exports/diff-review; the initialized database remains under .mochilab/diff-review. Remove these folders when no longer needed.

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
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • 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
Findings (7)

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
90% confidence
Finding

The reported general-purpose HTTP functionality substantially exceeds the advertised local comparison and manifest workflow, creating an unexpected path for private-file exfiltration or state-changing requests to external and internal services. Caller-enabled private-network access increases this exposure, while the documented approval requirement reduces risk but does not demonstrate technical enforcement; the helper source was not supplied, so reachability and safeguards remain unverified.

Content

No source excerpt is available for this finding.

Credential Access

High
Category
Privilege Escalation
Confidence
60% confidence
Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · tests/smoke.py (reported line 22)May include surrounding context.

python
class CoreTests(Base):
 def test_traversal(self):
  for name in ('../escape','a/../../escape','/etc/passwd','C:/Windows/test'):
   with self.assertRaises(ValueError):local(self.root,name)
 def test_existing_output_refused(self):
  p=self.file('one.txt','old')

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.

The runtime implements outbound HTTP requests, including remote mutations, unrelated to file comparison and delivery verification.

Medium
Category
Excessive Agency
Confidence
98% confidence
Finding

The safe_request helper accepts caller-supplied URLs, headers and bodies and supports POST, PUT, PATCH and DELETE (L090-L093), then opens a network connection and sends the request (L097-L108). This is an implemented network capability beyond the manifest's file comparison, delivery manifest and hash-verification purpose, although this file does not show it being invoked. HTTPS and address restrictions constrain destinations but do not justify the capability for this skill.

Content

No source excerpt is available for this finding.

HTTP requests can transmit supplied data without an accompanying disclosure.

Medium
Category
Output Handling
Confidence
86% confidence
Finding

safe_request sends caller-provided headers and request bodies to a supplied destination without a confirmation, operation-specific log, or explanatory docstring/comment about data transmission. HTTPS and address validation protect the connection but do not disclose that supplied data may leave the system.

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.

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

Low
Category
Output Handling
Confidence
90% confidence
Finding

Store initialization creates a .mochilab directory and SQLite database, and put persists supplied data there. Every runmain invocation initializes this store, but no confirmation, operation-specific log, or explanatory comment/docstring discloses that local state is retained.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.