ClawHub
Back to skill

Security audit

Privacy Redaction Desk

Security checks for vulnerabilities and agentic risk

Overview

Review is warranted because CSV copies can retain passwords and API keys while scan reports show no matches.

Do not rely on CSV exports being safe to share until secret-column masking is fixed. Inspect credential columns manually and review every output; a report with no matches does not prove sensitive information was removed. The inspected workflow showed no automatic uploads or malicious behavior.

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
  • 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
88% confidence
Finding

The reported general-purpose HTTP capabilities exceed the advertised local redaction workflow and could transmit sensitive source data or perform unauthorized remote mutations if misused. Access to explicitly opted-in private hosts could additionally expose internal services, although that opt-in requirement limits exposure and does not establish unrestricted SSRF. The instructions prohibit uploads and require host approval, but their effectiveness cannot be established without the helper implementation.

Content

No source excerpt is available for this finding.

CSV redaction does not redact secret fields based on their column names.

High
Category
Excessive Agency
Confidence
99% confidence
Finding

The CSV branch processes headers and cell values independently, never applying SENSITIVE to column names to redact their associated values. Consequently, arbitrary passwords, API keys, or other credentials in sensitive columns remain intact unless they happen to match another configured pattern. This is especially dangerous in a privacy-redaction tool because users may share the generated output believing those secrets were removed, while scan reports may omit them entirely.

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
76% confidence
Finding

The manifest declares no tool scope despite reported filesystem, shell, and network capabilities, creating a least-privilege gap for a skill that processes sensitive information. Local file access and Python execution fit the stated purpose, but unrestricted network access does not; workspace and approval instructions reduce risk but do not technically constrain the helper's capabilities. The practical impact depends on the host's permission enforcement and the implementation, which was not supplied.

Content

No source excerpt is available for this finding.

The redaction runtime includes an HTTP client capable of remote reads and mutations, unrelated to its stated file-redaction purpose.

Medium
Category
Excessive Agency
Confidence
96% confidence
Finding

The manifest describes scanning and redacting text, CSV, and JSON while preserving originals. The safe_request helper at L090–L115 instead implements outbound HTTP requests, including POST, PUT, PATCH, and DELETE with caller-supplied headers and bodies; this capability is not justified by that purpose. No invocation is shown in this file, so this finding concerns the exposed capability rather than demonstrated transmission of sensitive data.

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.