T08 · Insecure Dependencies
Warning
- Location
- SKILL.md:15
- Finding
- Unpinned Third-Party Packages Are Downloaded and Executed## Vulnerability Details **File Location**: `SKILL.md`, lines 15–17, 56, 63–67, 112–117, and 238–239 **Vulnerability Type**: Unpinned executable dependencies and unsafe supply-chain trust **Risk Level**: Medium ### Vulnerable Code `SKILL.md`, lines 15–17: ```bash npx hyperframes tts "Text here" --voice af_nova --output narration.wav npx hyperframes tts script.txt --voice bf_emma --output narration.wav npx hyperframes tts --list ``` `SKILL.md`, line 56: ```bash pip install kokoro-onnx soundfile ``` `SKILL.md`, lines 63–67: ```bash npx hyperframes transcribe audio.mp3 npx hyperframes transcribe video.mp4 --model small --language es npx hyperframes transcribe subtitles.srt npx hyperframes transcribe subtitles.vtt npx hyperframes transcribe openai-response.json ``` `SKILL.md`, lines 112–117: ```bash npx hyperframes remove-background subject.mp4 -o transparent.webm npx hyperframes remove-background subject.mp4 -o transparent.mov npx hyperframes remove-background portrait.jpg -o cutout.png npx hyperframes remove-background subject.mp4 -o subject.webm \ --background-output plate.webm npx hyperframes remove-background subject.mp4 -o transparent.webm --device cpu npx hyperframes remove-background --info ``` `SKILL.md`, lines 238–239: ```bash npx hyperframes tts script.txt --voice af_heart --output narration.wav npx hyperframes transcribe narration.wav ``` ### Technical Analysis The Skill repeatedly instructs users or agents to execute `hyperframes` through `npx` without specifying an audited package version. If the package is absent locally, `npx` can retrieve a currently available registry release and execute its lifecycle and application code with the invoking user's privileges. The Python dependencies `kokoro-onnx` and `soundfile` are also installed without version constraints or cryptographic hashes. The project contains no lockfile, dependency manifest with integrity meta ...[truncated 2111 chars]
- Remediation
- ## Remediation Suggestions 1. Pin `hyperframes` to a specific audited version in every command, such as: ```bash npx --yes hyperframes@1.2.3 tts "Text here" --voice af_nova --output narration.wav ``` 2. Maintain an npm lockfile containing package versions and integrity metadata. Prefer a reviewed local installation invoked through a package script rather than downloading executable code ad hoc on every use. 3. Pin Python dependencies in a requirements file: ```text kokoro-onnx==AUDITED_VERSION soundfile==AUDITED_VERSION ``` 4. Generate and enforce cryptographic hashes for Python dependencies, then install them with hash verification: ```bash python -m pip install --require-hashes -r requirements.txt ``` 5. Review and pin transitive dependencies where feasible. Update versions through a controlled process that includes source review, vulnerability scanning, and functional testing. 6. Document the approved npm and Python package indexes. Prevent fallback to untrusted mirrors and avoid adding dependency sources based on untrusted input. 7. Publish checksums and trusted download origins for the machine-learning model files downloaded on first use, and verify those checksums before loading the models. 8. Run media-processing dependencies in a least-privileged environment or container with restricted filesystem access, no unnecessary credentials, and limited outbound network connectivity.