The Agent Skills Directory

[SAFE]: The skill uses specialized MCP tools to interact with App Store Connect in a read-only manner. It processes information to update local project documentation such as ROADMAP.md and .planning/SIGNALS.md.
[PROMPT_INJECTION]: The skill processes untrusted external data in the form of App Store reviews and beta feedback. While this presents a surface for indirect prompt injection, the skill includes guardrails such as human-in-the-loop verification and a read-only stance towards the external API. * Ingestion points: Data fetched via mcp__asc-metadata__list_reviews and mcp__asc-metadata__list_beta_feedback_crashes in SKILL.md. * Boundary markers: The instructions do not define specific delimiters for separating user reviews from instructions, though they emphasize analysis and clustering. * Capability inventory: The skill has access to file system tools (Read, Write, Glob, Grep) to manage the local backlog. * Sanitization: No explicit sanitization of the review text is documented, but the agent's instructions are focused on extraction and normalization rather than direct execution of the content.

store-signals