The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes the gws command-line utility to perform email operations.
Evidence: Usage of gws gmail +send with various flags in SKILL.md.
[INDIRECT_PROMPT_INJECTION]: The skill provides a surface where untrusted data could be used to construct email content, potentially leading to unintended actions if the agent follows instructions embedded in that data.
Ingestion points: The --subject and --body parameters in the gws command in SKILL.md are ingestion points for potentially untrusted content.
Boundary markers: Absent; there are no delimiters or specific instructions to the agent to treat input data as literal content rather than instructions.
Capability inventory: The skill has the capability to send data to external recipients via the gws binary.
Sanitization: No input sanitization or validation logic is specified in the skill instructions.

gws-gmail-send