Back to skill

Security audit

CorelDRAW Editor

Security checks for vulnerabilities and agentic risk

Overview

This skill automates CorelDRAW file inspection, export, and user-directed edits, with destructive document changes disclosed and guarded by copy-first guidance and overwrite flags.

Install only in a Windows environment where you trust the Python package source and have CorelDRAW installed. Use copies for edits, avoid --in-place unless you intentionally want to modify the original CDR, and only use --overwrite when replacing an existing output is intended.

Vulnerability Patterns
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • 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
Findings (1)

T08 · Insecure Dependencies

Warning
Location
SKILL.md:14
Finding

Unpinned Third-Party Dependency Installation

Content
View full analysis

Vulnerability Details

File Location: SKILL.md, line 14
Vulnerability Type: Unpinned dependency installation from the configured Python package index
Risk Level: Medium

Complete Code Snippet:

markdown
- `pywin32`. If missing and local package installation is acceptable, run `py -3 -m pip install --user pywin32`.

Technical Analysis

The documented installation command requests pywin32 without specifying a reviewed version or validating the downloaded artifact with a cryptographic hash. Consequently, the actual package installed depends on the state of the configured Python package index at installation time.

If the package source, index configuration, distribution account, or delivery channel is compromised, a user following this instruction could install attacker-controlled code. Python packages may execute installation or build logic, and their imported modules subsequently execute with the privileges of the invoking user.

This finding is limited to dependency integrity. The audited project does not itself retrieve remote payloads, execute shell commands, establish persistence, or collect credentials.

Attack Path

  1. An attacker compromises the package distribution channel, a package publisher account, or the Python index configured on the target system.
  2. The attacker makes a malicious pywin32 artifact available as the version selected by the unpinned command.
  3. A user follows the documented command: py -3 -m pip install --user pywin32.
  4. Pip downloads and installs the attacker-controlled artifact without checking a project-specified version or hash.
  5. Malicious installation logic or imported package code executes under the invoking user's account.

Exploitation therefore requires control of, or successful redirection of, the dependency supply channel used by the target.

Impact Assessment

Successful exploitation could provide arbitrary code execution with the ...[truncated 418 chars]

Remediation
View remediation

Remediation Suggestions

  • Pin pywin32 to a specifically reviewed version instead of resolving the latest available release.
  • Store dependency declarations in a locked requirements file and include cryptographic hashes for every accepted distribution.
  • Install with hash verification, for example:
    text
    py -3 -m pip install --require-hashes -r requirements.txt
    
  • Generate hashes from artifacts obtained through a trusted release process and review updates before changing the lock file.
  • Where organizational policy permits, specify an approved package index and prevent fallback to untrusted or dependency-confusion-prone sources.
  • Prefer a managed virtual environment over modification of the user's general Python package environment.
  • Document that users should verify the configured pip index and avoid installing the dependency when package-source integrity cannot be established.
Vulnerability Patterns
  • Rogue AgentSelf-Modification, Session Persistence
  • Behavioral ASTexec() Call, eval() Call, Dynamic Import
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Findings (5)

Self-Modification

High
Category
Rogue Agent
Confidence
85% confidence
Finding

Skill modifies its own code, configuration, or behavior at runtime. Self-modification enables an agent to escalate privileges, disable safety constraints, or install persistent backdoors.

Content

Scanner excerpt · scripts/coreldraw_editor.py (reported line 195)May include surrounding context.

python
for filename, indexes in items:
            out_path = out_dir / filename
            if out_path.exists() and not args.overwrite:
                raise FileExistsError(f"Refusing to overwrite existing file: {out_path}")
            create_selection(layer, indexes)
            if args.format in {"png", "jpg", "jpeg"}:
                export_bitmap(doc, out_path, args.format, args.dpi, args.transparent)

Self-Modification

High
Category
Rogue Agent
Confidence
85% confidence
Finding

Skill modifies its own code, configuration, or behavior at runtime. Self-modification enables an agent to escalate privileges, disable safety constraints, or install persistent backdoors.

Content

Scanner excerpt · scripts/coreldraw_editor.py (reported line 300)May include surrounding context.

python
for filename, indexes in items:
            out_path = out_dir / filename
            if out_path.exists() and not args.overwrite:
                raise FileExistsError(f"Refusing to overwrite existing file: {out_path}")
            create_selection(layer, indexes)
            if args.format in {"png", "jpg", "jpeg"}:
                export_bitmap(doc, out_path, args.format, args.dpi, args.transparent)

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
89% confidence
Finding

The skill describes file inspection, editing, export, and overwrite-capable workflows but does not declare any explicit tool scope such as permissions or allowed-tools. In an agent environment, that omission can lead to broader-than-necessary file read/write access, increasing the risk of unintended modification or exfiltration of local files if the skill is invoked on attacker-controlled paths or with ambiguous instructions.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

This code accepts plan operations including delete, move, resize, rotate, ungroup, and layer visibility/printability changes, then saves the modified document in place or to a new file. Although the operations are part of the tool's editing purpose, there is no user-facing warning or confirmation specifically disclosing that apply-plan can make destructive or irreversible document changes, especially when --in-place is used.

Content

No source excerpt is available for this finding.

Dynamic attribute access via getattr()

Low
Category
Dangerous Code Execution
Confidence
50% confidence
Finding

Dynamic getattr() with a non-literal attribute name can access arbitrary object attributes, potentially bypassing access controls.

Content

Scanner excerpt · scripts/coreldraw_editor.py (reported line 54)May include surrounding context.

python
def const_value(name: str, fallback: int | None = None) -> int:
    try:
        return int(getattr(win32com.client.constants, name))
    except Exception:
        if fallback is None:
            raise