The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses various shell commands such as open, lsof, and osascript to launch and manage desktop applications. It includes specific safety guidelines to prevent the agent from accidentally terminating its own process during these operations.
[DYNAMIC_EXECUTION]: The skill utilizes playwright-cli eval and run-code to execute scripts within the context of the target Electron application. This is a core feature required for the skill's primary purpose of UI automation and state validation.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external applications, which creates a surface for indirect prompt injection.
Ingestion points: UI snapshots and metadata are ingested via playwright-cli and curl requests to the application's debugging port (SKILL.md).
Boundary markers: The instructions do not define delimiters or provide warnings to ignore instructions found within the application UI.
Capability inventory: The skill has access to shell execution (Bash), process management, and file system tools (Read, Grep).
Sanitization: There is no evidence of sanitization for the content retrieved from the automated applications before it is processed by the agent.
[REMOTE_CODE_EXECUTION]: Automated scans flagged a pattern where curl output is piped to python3. This finding is a false positive; the instruction uses python3 -c with a static, hardcoded script to format JSON data from a local port (localhost:9222), rather than executing remote code.

electron