The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes content from the local codebase to identify patterns and propose standards. While reading untrusted data (source code) presents a theoretical surface for indirect prompt injection, the risk is low because the skill's actions are focused on summarizing observations and writing them to a documentation file under user supervision.
[DATA_EXPOSURE]: To perform its function, the skill scans local project files and directories. This file access is legitimate and required for the skill's stated purpose of extracting coding conventions. No evidence of credential harvesting or external data transmission was found.

grill-me-standards