The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill defines workflows like /phx:review and /phx:investigate that ingest untrusted user code while recommending high-privilege tools. 1. Ingestion points: Workflow definitions in SKILL.md for code review and bug investigation. 2. Boundary markers: Absent; there are no instructions to delimit or distrust input code or error logs. 3. Capability inventory: Utilizes mcp__tidewave__project_eval (runtime code evaluation) and mcp__tidewave__execute_sql_query (database access). 4. Sanitization: Absent; no input validation or filtering is specified before tool usage.
[DYNAMIC_EXECUTION]: The skill utilizes the mcp__tidewave__project_eval tool to execute Elixir code directly within the running application environment for debugging purposes.
[COMMAND_EXECUTION]: The skill employs mcp__tidewave__execute_sql_query to run SQL commands against the database for verifying state during development workflows.
[EXTERNAL_DOWNLOADS]: The skill references documentation and guides from hexdocs.pm, which is the official and well-known hosting service for Elixir and Phoenix package documentation.

examples