The Agent Skills Directory

[COMMAND_EXECUTION]: The skill includes multiple code examples in SQL, CQL, and JavaScript. These are provided as static reference materials to assist users in understanding database modeling. There are no instructions or capabilities that would lead to the execution of these commands on the host system.
[EXTERNAL_DOWNLOADS]: The skill does not initiate any network connections or downloads. All reference documentation and example files are contained within the skill's local directory.
[DATA_EXFILTRATION]: The skill does not access sensitive files, environment variables, or system credentials. All data used in examples (such as UUIDs, names, and product info) is placeholder dummy data.
[PROMPT_INJECTION]: The instructions in SKILL.md provide a structured workflow for the agent to diagnose and resolve normalization issues using provided reference files. There are no attempts to override agent behavior or bypass safety guidelines.
[OBFUSCATION]: No obfuscated content, encoded strings (beyond standard UUID formats), or hidden instructions were found in any of the 58 analyzed files.
[INDIRECT_PROMPT_INJECTION]: While the skill processes user-provided database schemas, it lacks exploitable capabilities like file writing or subprocess execution. The skill acts purely as a knowledge base, presenting no significant surface for indirect injection attacks.

database-normalization