The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the Bash tool to execute adk CLI commands (e.g., search, info, list). These commands are constructed using user-supplied strings, representing a potential command injection surface if the inputs are not properly sanitized by the platform's tool executor.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted user input to perform searches and execute commands without explicit boundary markers or sanitization guidelines.
Ingestion points: User queries processed in SKILL.md regarding ADK concepts and integration details.
Boundary markers: None found in the instructions for tool usage.
Capability inventory: Bash (adk CLI), Glob, Grep, and Read tools as described in SKILL.md.
Sanitization: None described for the interpolation of user strings into shell commands or search patterns.
[EXTERNAL_DOWNLOADS]: The skill documents the use of adk add to install integrations. This action fetches and registers external code from the Botpress ecosystem, which is a standard feature for this well-known development framework.

adk