The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the vendor-provided postplus command-line interface to execute search-intent research tasks, perform diagnostic checks (postplus doctor), and manage software updates. These operations are consistent with the skill's stated purpose of conducting marketplace research.
[DATA_EXPOSURE]: Research data and watchlist caches are managed within a dedicated local directory (.postplus/google-trends/), which follows best practices for local storage and allows for user oversight of processed information.
[INDIRECT_PROMPT_INJECTION]: The skill has a data ingestion surface as it processes external search trend results.
Ingestion points: External data is collected into ./result.json via the google-trends-fast command.
Boundary markers: The instructions explicitly require the agent to "separate observation from inference" and validate records against a specific schema defined in the vendor's shared reference files.
Capability inventory: The agent is restricted to postplus CLI subcommands and does not have arbitrary code execution capabilities.
Sanitization: The workflow involves filtering for specific "trend signals" and normalized outputs, reducing the risk of processing malicious metadata embedded in search trends.
[SAFE]: The authentication flow (postplus auth login) is designed to be user-interactive, requiring the agent to provide a URL for the user to visit rather than attempting to handle credentials or tokens automatically.

google-trends-research