Safety Report
Full AI pipeline to create dark motivational TikTok/Reels videos using REAL video footage. Generates script (Claude), voiceover (ElevenLabs), searches real d...
Security Analysis
medium confidenceThe skill's code and instructions match its stated purpose (creating TikTok-style videos from Pexels + LLM/voice APIs) but the package metadata omits required API environment variables and a system-prompt-override pattern was detected — review API key usage and run in an isolated environment before use.
The skill's stated goal (script via Anthropic Claude, voice via ElevenLabs, footage via Pexels, composition with MoviePy/FFmpeg) aligns with the included code and SKILL.md. However, the registry metadata lists no required env vars while the code expects ANTHROPIC_API_KEY, ELEVENLABS_API_KEY, and PEXELS_API_KEY — an inconsistency between declared requirements and actual needs.
The SKILL.md and pipeline_full.py instruct the agent to call Anthropic, ElevenLabs, and Pexels APIs, download video files from Pexels, produce audio, and compose a vertical MP4. All actions are within the stated video-creation scope and do not attempt to read unrelated system files or exfiltrate data to unknown endpoints.
There is no install spec (instruction-only plus an included Python script). The pipeline requires standard Python packages and FFmpeg but does not perform any opaque archive downloads or install arbitrary binaries from untrusted URLs.
The code reads three API keys from environment variables (ANTHROPIC_API_KEY, ELEVENLABS_API_KEY, PEXELS_API_KEY) but the skill metadata declares no required env vars or primary credential. This mismatch is potentially misleading: users will need to supply keys, and the metadata should explicitly list them. Otherwise, the requested credentials are proportional to the task (each key is used for the corresponding API).
The skill is not always-enabled and does not request elevated platform privileges. It does write downloaded clips and generated files to disk (expected for a video pipeline) but does not modify other skills or system-wide agent settings.
Guidance
Before installing: (1) Expect to provide three API keys (Anthropic/Claude, ElevenLabs, Pexels) even though the package metadata doesn't list them — verify and store them securely (use least-privilege keys if possible). (2) Inspect pipeline_full.py (we reviewed it) and confirm you are comfortable with it downloading and writing video/audio files to the local filesystem; run it in an isolated environment (container or VM) if you don't fully trust it. (3) The SKILL.md includes a system prompt to force JSON output from Claude — this is normal for parsing but should be audited to ensure it doesn't attempt to override the agent's broader system prompts. (4) Ensure you have legal/usage rights for Pexels content and ElevenLabs voice usage, and install FFmpeg and the recommended font safely. (5) If you plan to allow autonomous agent invocation, consider the blast radius of the API keys you supply; otherwise run manually. If the registry metadata cannot be corrected to list the required env vars, treat the omission as a red flag and obtain clarification from the publisher before use.
Latest Release
v1.0.0
tiktok-motivation-video skill v1.0.0 - Initial release: Full pipeline to generate dark motivational TikTok/Reels videos using real video footage (no AI visuals). - Creates script with precise timing and Pexels search keywords via Claude API. - Generates voiceover using ElevenLabs (supports custom voice, English/Vietnamese). - Fetches high-quality cinematic vertical videos from Pexels API. - Composes video with animated text overlays (MoviePy) and applies color grading (FFmpeg). - Outputs ready-to-post 1080x1920 MP4 with TikTok-compliant formatting and captions.
Popular Skills
Published by @thuongvh2-python on ClawHub
