Back to skill

Security audit

Claw Presenter

Security checks for vulnerabilities and agentic risk

Overview

This skill does what it claims: it converts user-provided PPT/PDF files into slide images and narration data, with ordinary install and local-output cautions.

Install this in a virtual environment or sandbox if possible, review before allowing sudo package installs, and avoid processing untrusted or highly sensitive decks unless you are comfortable with their extracted notes and slide text being saved locally under presentations/<name>.

Vulnerability Patterns
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (7)

Vague Triggers

Medium
Confidence
84% confidence
Finding
The trigger phrases are broad enough to match ordinary requests for presentation help, which can cause this skill to activate when the user did not intend file parsing or shell-backed processing. In context, that matters because the skill may then ask for or operate on local files and generate outputs, expanding access beyond a simple conversational request.

Missing User Warnings

Medium
Confidence
93% confidence
Finding
The script writes extracted slide text, speaker notes, and generated metadata to presentation.json under a persistent output directory without any explicit warning, consent, retention control, or cleanup. In this skill's context, presentation notes commonly contain sensitive internal speaking points, unpublished content, or secrets, so silent persistence increases the risk of unintended disclosure to other users, tools, or later processes.

Unpinned Dependencies

Low
Category
Supply Chain
Content
python-pptx
Pillow
pdf2image
pdfplumber
Confidence
96% confidence
Finding
The dependency python-pptx is unpinned, so installs may resolve to different versions over time. This creates supply-chain and reproducibility risk because a future release could introduce security regressions or breaking changes without review.

Unpinned Dependencies

Low
Category
Supply Chain
Content
python-pptx
Pillow
pdf2image
pdfplumber
Confidence
99% confidence
Finding
Pillow is unpinned, allowing installation of any available version, including versions with known vulnerabilities. In this skill, the library is used for image processing of PPT/PDF-derived content, so malformed or attacker-supplied files could trigger denial of service or worse if a vulnerable version is installed.

Unpinned Dependencies

Low
Category
Supply Chain
Content
python-pptx
Pillow
pdf2image
pdfplumber
Confidence
95% confidence
Finding
The dependency pdf2image is unpinned, which weakens build reproducibility and can expose the skill to future vulnerable releases or dependency confusion in the supply chain. Because this skill processes external presentation documents, dependency integrity matters.

Unpinned Dependencies

Low
Category
Supply Chain
Content
python-pptx
Pillow
pdf2image
pdfplumber
Confidence
95% confidence
Finding
The dependency pdfplumber is unpinned, creating supply-chain risk and non-deterministic environments. Since the skill parses PDF content that may come from users, relying on uncontrolled package versions increases exposure to parser bugs and future CVEs.

Known Vulnerable Dependency: Pillow — 10 advisory(ies): CVE-2016-2533 (Pillow buffer overflow in ImagingPcdDecode); CVE-2023-50447 (Arbitrary Code Execution in Pillow); CVE-2021-27922 (Pillow Uncontrolled Resource Consumption) +7 more

Critical
Category
Supply Chain
Confidence
99% confidence
Finding
The requirements file includes Pillow without a version constraint, and the scanner reports multiple known advisories affecting Pillow. In a skill that converts slides and documents into images, attacker-controlled PPT/PDF-derived image content may be processed by Pillow, increasing the likelihood of exploiting image parsing flaws for denial of service or potentially arbitrary code execution depending on the resolved version.

Static analysis

No suspicious patterns detected.