The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the run_agents tool for sub-agent orchestration and suggests using version control commands like git worktree and git branch to ensure isolation when code changes are allowed.- [PROMPT_INJECTION]: The skill identifies a surface for indirect prompt injection by processing untrusted data from sources such as pull requests, issue descriptions, and codebase artifacts.
Ingestion points: Untrusted data enters the context when investigating pull requests, issues, or codebase files as described in steps 1 and 3.
Boundary markers: The instructions do not explicitly require the use of delimiters or specific warnings to ignore embedded instructions within the artifacts being analyzed.
Capability inventory: The skill uses run_agents to launch sub-agents that may have the ability to modify local files or execute shell commands.
Sanitization: There are no explicit steps for sanitizing or escaping the content of the analyzed artifacts before providing them to sub-agents.

council