The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands using the connector-status.py script, specifically python "${CLAUDE_PLUGIN_ROOT}/scripts/connector-status.py" --action setup-guide --name <connector>. The <connector> placeholder is directly populated by user-supplied input, creating a command injection vulnerability if the agent does not properly sanitize the input.
[INDIRECT_PROMPT_INJECTION]: The skill has a broad attack surface for indirect injection:
Ingestion points: The connector-name argument supplied by the user at runtime in SKILL.md.
Boundary markers: None provided; the skill does not instruct the agent to validate or delimit the user input before processing.
Capability inventory: The skill has the capability to execute shell commands via Python as seen in the Process section of SKILL.md.
Sanitization: The instructions lack any mention of sanitizing or escaping the <connector> placeholder before it is interpolated into the shell command.
[EXTERNAL_DOWNLOADS]: The skill references several official MCP server endpoints for setup, including:
https://mcp.facebook.com/ads for Meta Ads.
https://mcp.hubspot.com for HubSpot.
Mentions of Amazon Ads and Google Ads specific integration paths.
These are documented as official connection targets for user configuration.

connect