The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the agent to download and install a CLI tool using npx from @mvanhorn/printing-press-library and go install from github.com/mvanhorn/printing-press-library. These resources are owned by the skill's verified vendor and are used for the primary purpose of the skill.
[COMMAND_EXECUTION]: The skill relies heavily on the execution of shell commands through the sendgrid-pp-cli binary, including installation commands and complex management tasks.
[DATA_EXFILTRATION]: The CLI tool includes a built-in --deliver webhook:<url> feature that allows the agent to POST command results directly to any external URL. While intended for automation, this provides a direct channel for exfiltrating sensitive SendGrid data or customer PII if the URL parameter is manipulated.
[INDIRECT_PROMPT_INJECTION]: The skill possesses a significant attack surface for indirect prompt injection as it ingests and processes untrusted data from external sources.
Ingestion points: The skill reads data from CSV files (suppression sync --from, suppression diff --against) and JSON payloads (templates lint --against).
Boundary markers: No explicit instructions are provided to the agent regarding the use of delimiters or warnings to ignore instructions embedded within these external files.
Capability inventory: The skill has the capability to write to the local filesystem (--deliver file:<path>, SQLite mirror) and transmit data over the network via the webhook delivery sink and the feedback command.
Sanitization: There is no mention of sanitization or validation of the content within the ingested CSV or JSON files before they are processed by the CLI or agent.

pp-sendgrid