The Agent Skills Directory

[COMMAND_EXECUTION]: The skill is configured with the Bash tool, allowing it to execute shell commands. While no malicious commands are hardcoded, this capability allows the agent to perform system-level operations based on user or external input.
[PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection by ingesting untrusted data through WebSearch and Read tools while possessing write and execution capabilities.
Ingestion points: Data is ingested via WebSearch, Read, Glob, and Grep tools (SKILL.md).
Boundary markers: No explicit delimiters or instructions are provided to the model to ignore embedded commands within the ingested data.
Capability inventory: The skill has access to Bash (command execution), Write, and Edit (file system modifications).
Sanitization: There are no instructions for sanitizing or validating content retrieved from external sources before processing.

product-manager