The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes a local Python script (scripts/fetch_similar.py) using python3. The script's logic is transparent and relies exclusively on the Python standard library to interact with the vendor's API.
[EXTERNAL_DOWNLOADS]: The skill makes HTTP POST requests to doubaoya.com to retrieve account data and submit synchronization requests. This domain is the official service provider for the skill, and the communication is essential for the stated functionality.
[DATA_EXFILTRATION]: While the skill sends account names and an API key to doubaoya.com, this is the intended operation. The API key is securely handled via the DOUBAOYA_API_KEY environment variable and is not exposed in logs or returned to the user.
[INDIRECT_PROMPT_INJECTION]: The skill processes data from an external API which is then presented to the agent.
Ingestion points: API response data fields (accountName, avgReadCount, similarity) processed in scripts/fetch_similar.py and SKILL.md.
Boundary markers: The instructions in SKILL.md direct the agent to perform 'defensive reading' and map specific fields into a structured Markdown table.
Capability inventory: The skill possesses network request capabilities via a script; no file-writing or arbitrary command execution capabilities are granted to the processing of the external data.
Sanitization: The script outputs structured JSON, and the agent is instructed to extract only defined fields, minimizing the risk of executing embedded instructions.

wechat-similar-account