The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses shell-based helper functions to execute redis-cli, python3, awk, and grep for server discovery, performance monitoring, and slow log analysis.
[CREDENTIALS_UNSAFE]: The implementation uses the -a flag with redis-cli to provide the database password. This approach can leak credentials to other system users through process monitoring tools like ps. The skill documentation explicitly flags this as a pitfall and suggests alternative environment variables.
[PROMPT_INJECTION]: The skill is subject to indirect prompt injection risks due to its data ingestion surface.
Ingestion points: Data retrieved from Redis (e.g., key names, configuration values, slow log entries, and client lists) enters the agent's context and is processed by scripts in SKILL.md.
Boundary markers: The skill does not employ specific delimiters or instructions to prevent the agent from being influenced by data returned from the Redis server.
Capability inventory: The skill has the capability to execute shell commands and Python code locally based on data processed from Redis.
Sanitization: There is no explicit sanitization of the values or keys retrieved from the database before they are used in scripts or presented to the agent.

managing-redis