The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill is configured to automatically clone a third-party repository from https://github.com/tmwgsicp/wechat-download-api.git into the skill directory if it is not already present.
[REMOTE_CODE_EXECUTION]: The shell script scripts/start_wechat_api.sh executes python3 app.py from the externally cloned repository, allowing the execution of unverified code on the host machine.
[COMMAND_EXECUTION]: The skill performs several command-line operations, including installing Python packages via pip3 install, managing background processes with nohup and kill, and using git clone to fetch external assets.
[INDIRECT_PROMPT_INJECTION]: The skill presents an attack surface for indirect prompt injection as it processes and stores untrusted article content from WeChat.
Ingestion points: Article data is fetched from the local API service by scripts/download_wechat.py and written to local files.
Boundary markers: No specific delimiters or safety instructions are used to isolate the ingested external content.
Capability inventory: The skill has the ability to clone repositories, install packages, execute Python scripts, and write to the local filesystem.
Sanitization: Filenames are sanitized using regex to remove dangerous characters, but the article body is saved as raw text without validation or filtering.

wechat-download