The Agent Skills Directory

[SAFE] (SAFE): No malicious patterns, prompt injections, or data exfiltration attempts were found during the analysis of the code and metadata.
[COMMAND_EXECUTION] (LOW): The skill utilizes Redis Lua scripts (eval) and PostgreSQL advisory locks (session.execute). These are implemented as hardcoded strings with parameterized inputs, which is the industry-standard approach for preventing injection.
[EXTERNAL_DOWNLOADS] (LOW): The code references standard libraries and reputable packages such as redis, sqlalchemy, and uuid_utils. These dependencies are appropriate for the skill's functional requirements.
[INDIRECT_PROMPT_INJECTION] (LOW): The skill handles resource identifiers provided at runtime. Because it uses cryptographic hashing (MD5) and robust database parameterization, these inputs cannot be used to inject malicious commands or alter the logic of the lock backends.

distributed-locks