The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill's primary function is to read and analyze external Rails codebases to identify architectural patterns and violations. This ingestion of untrusted data creates a surface for indirect prompt injection, where malicious instructions embedded in the analyzed code (such as in comments or markdown files) could attempt to influence the agent's output or actions.
Ingestion points: Workflows like workflows/analyze.md and workflows/review.md read arbitrary files from the user's project directory to perform audits.
Capability inventory: The skill uses Grep, Glob, Read, and Task tools to examine and process file contents.
[SAFE]: The skill primarily serves as a knowledge base and a set of analysis procedures for Rails development. It references several Ruby gems, including action_policy, active_delivery, anyway_config, and rubanok, which are established open-source libraries maintained by the author. All shell commands used for analysis (e.g., find, grep, wc, git log) are standard utilities focused on metric collection and code inspection. No evidence of malicious code execution, data exfiltration, or persistence mechanisms was detected.

layered-rails