The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill documentation instructs the user to install the opencode-ai package globally using npm (npm install -g opencode-ai@latest).
[COMMAND_EXECUTION]: The skill executes several shell commands to interact with the OpenCode tool, including opencode run, opencode auth list, and opencode --version. It also runs a local Python script using uv run python scripts/delegation_executor.py.
[INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied prompts and passes them to the opencode command-line interface, creating a potential vector for indirect injection.
Ingestion points: The PROMPT argument in the Makefile example and positional arguments in the opencode run command (SKILL.md).
Boundary markers: No delimiters or 'ignore' instructions are visible in the command construction examples to separate user input from the execution context.
Capability inventory: The skill has the ability to execute the opencode binary, which can interact with various AI models and perform delegated tasks (SKILL.md).
Sanitization: There is no evidence of prompt sanitization, escaping, or validation before the user input is passed to the external tool.

opencode-delegation