The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute standard local development commands such as mix compile, mix ecto.migrate, and mix test to diagnose the state of the application.
[INDIRECT_PROMPT_INJECTION]: The skill operates on untrusted data from bug descriptions and application logs, creating a potential surface for indirect prompt injection. Ingestion points: User-supplied $ARGUMENTS and runtime logs retrieved via mcp__tidewave__get_logs in SKILL.md. Boundary markers: The instructions do not define specific delimiters or guardrails to isolate ingested data from the agent's control logic. Capability inventory: The agent can perform database queries and code evaluation through mcp__tidewave__execute_sql_query and mcp__tidewave__project_eval, and run shell commands via mix and grep. Sanitization: There is no explicit sanitization or filtering of the content retrieved from external log files before it is processed by the agent.
[DYNAMIC_EXECUTION]: The skill utilizes mcp__tidewave__project_eval to evaluate Elixir code snippets at runtime to test debugging hypotheses.

investigate