The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill retrieves article content, follower nicknames, and metadata from the WeChat MP API. This data constitutes an untrusted input surface that could potentially contain hidden instructions designed to influence the agent's logic.
Ingestion points: The draft/get, freepublish/batchget, and user/info/batchget recipes in SKILL.md ingest external data.
Boundary markers: There are no explicit delimiters or warnings to the agent to disregard instructions within the retrieved content.
Capability inventory: The skill utilizes the Bash tool, which allows for network requests and file system interactions.
Sanitization: The skill employs jq to ensure safe shell interpolation when sending data to the API, which prevents command injection. However, it does not sanitize the content returned by the API before it is processed by the agent.
[COMMAND_EXECUTION]: The skill relies on the Bash tool to perform operations using curl and jq. It manages the WeChat authentication flow by exchanging environment-provided credentials for a temporary access token. The access token is cached in a file within the /tmp directory, which identifies the application ID in the filename.
[EXTERNAL_DOWNLOADS]: The skill fetches data and configuration from the official WeChat API (api.weixin.qq.com). These network operations are standard for the skill's intended purpose of interacting with the WeChat messaging platform.

wechat-official-account