The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructions involve executing the @useclawlink/cli tool via npx to authenticate and perform Slack operations.
[PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface by ingesting Slack conversation data which may contain instructions from external participants.
Ingestion points: Slack conversation history, threads, and search results retrieved through actions such as slack_fetch_conversation_history and slack_search_messages in SKILL.md.
Boundary markers: The documentation instructs the agent to "confirm with the user before any write," acting as a human-in-the-loop validation step.
Capability inventory: The skill includes various high-privilege Slack actions such as sending messages (slack_send_message), managing channels (slack_create_channel), and file operations (slack_upload_or_create_a_file_in_slack) performed via the CLI.
Sanitization: There are no explicit sanitization or filtering instructions defined for the retrieved Slack message strings within the skill documentation.

slack