The Agent Skills Directory

[DATA_EXPOSURE]: The skill reads from a local file named app-marketing-context.md to gain context about the user's project. This is a legitimate practice for maintaining state and context within the agent environment.
[COMMAND_EXECUTION]: The skill references several specialized tools such as search_apps, get_app_detail, and batch_keyword_difficulty. These represent structured tool calls intended for the agent's execution environment to interface with market data APIs.
[PROMPT_INJECTION]: The skill is subject to indirect prompt injection risks as it processes external content (titles, descriptions, and reviews) from competitor apps. While this is an inherent risk for any skill processing web/API data, no malicious behavior was found in the skill's own instructions.
Ingestion points: Data enters the context via app-marketing-context.md and competitor details fetched via get_app_detail (SKILL.md).
Boundary markers: None present.
Capability inventory: Uses search_apps, get_app_detail, and batch_keyword_difficulty for data retrieval; no dangerous file system or shell operations detected.
Sanitization: No explicit sanitization or filtering of fetched app metadata is mentioned.

competitor-analysis