Back to skill

Security audit

PyInstaller 打包 Windows exe

Security checks for vulnerabilities and agentic risk

Overview

ClawScan marks this skill suspicious because scanners report actions that could disrupt the desktop or overwrite files without warning, and analysis coverage is partial.

Review the Explorer repair and Desktop overwrite behavior before installing. The scan is incomplete: one workflow was only partly inspected, and the image was outside analyzer coverage, so the available results do not establish that every component is safe.

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
Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • Tool MisuseTool Parameter Abuse, Chaining Abuse, Unsafe Defaults
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Findings (13)

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
75% confidence
Finding

Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).

Content

Scanner excerpt · SKILL.md (reported line 27)May include surrounding context.

Core build command

bat
python -m PyInstaller --noconfirm --onefile --windowed ^
  --name appname ^
  --icon "absolute\path\to\icon.ico" ^
  --distpath dist --workpath build_tmp --specpath build_tmp ^

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
75% confidence
Finding

Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).

Content

Scanner excerpt · references/build-exe-cmd.md (reported line 22)May include surrounding context.

md
if %errorlevel%==0 (set "PY=python") else (set "PY=C:\Python313\python.exe")

echo [1/2] 打包 model_manager.exe ...
"%PY%" -m PyInstaller --noconfirm --onefile --windowed ^
  --name model_manager ^
  --icon "%BASE%mm.ico" ^
  --distpath "%BASE%dist" ^

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
75% confidence
Finding

Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).

Content

Scanner excerpt · references/build-exe-cmd.md (reported line 37)May include surrounding context.

md
)

echo [2/2] 打包 claude_switcher.exe ...
"%PY%" -m PyInstaller --noconfirm --onefile --windowed ^
  --name claude_switcher ^
  --icon "%BASE%mm.ico" ^
  --distpath "%BASE%dist" ^

Incomplete referenced artifact analysis

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 152)May include surrounding context.

md
`assets/build-exe.yml` is a ready workflow: runs on `windows-latest`, builds both

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
75% confidence
Finding

Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).

Content

Scanner excerpt · assets/build-exe.yml (reported line 50)May include surrounding context.

yaml
- name: Build model_manager.exe
        run: |
          python -m PyInstaller --noconfirm --onefile --windowed `
            --name model_manager `
            --icon "$env:GITHUB_WORKSPACE\mm.ico" `
            --distpath dist `

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
75% confidence
Finding

Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).

Content

Scanner excerpt · assets/build-exe.yml (reported line 60)May include surrounding context.

yaml
- name: Build claude_switcher.exe
        run: |
          python -m PyInstaller --noconfirm --onefile --windowed `
            --name claude_switcher `
            --icon "$env:GITHUB_WORKSPACE\mm.ico" `
            --distpath dist `

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
65% confidence
Finding

Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.

Content

Scanner excerpt · SKILL.md (reported line 27)May include surrounding context.

Core build command

bat
python -m PyInstaller --noconfirm --onefile --windowed ^
  --name appname ^
  --icon "absolute\path\to\icon.ico" ^
  --distpath dist --workpath build_tmp --specpath build_tmp ^

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
65% confidence
Finding

Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.

Content

Scanner excerpt · references/build-exe-cmd.md (reported line 22)May include surrounding context.

md
if %errorlevel%==0 (set "PY=python") else (set "PY=C:\Python313\python.exe")

echo [1/2] 打包 model_manager.exe ...
"%PY%" -m PyInstaller --noconfirm --onefile --windowed ^
  --name model_manager ^
  --icon "%BASE%mm.ico" ^
  --distpath "%BASE%dist" ^

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
65% confidence
Finding

Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.

Content

Scanner excerpt · references/build-exe-cmd.md (reported line 37)May include surrounding context.

md
)

echo [2/2] 打包 claude_switcher.exe ...
"%PY%" -m PyInstaller --noconfirm --onefile --windowed ^
  --name claude_switcher ^
  --icon "%BASE%mm.ico" ^
  --distpath "%BASE%dist" ^

The icon-cache repair forcibly terminates Windows Explorer without warning about disruption to the user's desktop session.

Medium
Category
Output Handling
Confidence
94% confidence
Finding

The repair forcibly terminates Explorer, disrupting the desktop shell and potentially interrupting Explorer-managed file operations. Icon-cache troubleshooting is relevant to this skill, but the instructions omit a disruption warning and do not ensure Explorer restarts if an intermediate operation fails.

Content

No source excerpt is available for this finding.

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
65% confidence
Finding

Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.

Content

Scanner excerpt · assets/build-exe.yml (reported line 50)May include surrounding context.

yaml
- name: Build model_manager.exe
        run: |
          python -m PyInstaller --noconfirm --onefile --windowed `
            --name model_manager `
            --icon "$env:GITHUB_WORKSPACE\mm.ico" `
            --distpath dist `

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
65% confidence
Finding

Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.

Content

Scanner excerpt · assets/build-exe.yml (reported line 60)May include surrounding context.

yaml
- name: Build claude_switcher.exe
        run: |
          python -m PyInstaller --noconfirm --onefile --windowed `
            --name claude_switcher `
            --icon "$env:GITHUB_WORKSPACE\mm.ico" `
            --distpath dist `

The build workflow silently overwrites existing desktop executables without warning users.

Medium
Category
Output Handling
Confidence
95% confidence
Finding

The copy /y commands unconditionally replace same-named executables on the user's Desktop, outside the dedicated build directory. Although Desktop delivery is an intended feature, an existing customized or unrelated executable can be lost without confirmation or backup.

Content

No source excerpt is available for this finding.