The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes various composio CLI commands within a Bash environment to interact with integrated services. This includes searching for tools, listing and fetching ticket data from Gorgias, and discovering Slack channels or sending messages. These actions are standard for the skill's intended purpose of automated reporting.
[SAFE]: The skill processes untrusted external data from support tickets, which presents a surface for indirect prompt injection.
Ingestion points: Ticket details and message content retrieved from Gorgias via the GORGIAS_GET_TICKET command in Step 3.
Boundary markers: There are no explicit delimiters or protective warnings in the markdown template used to compile the digest in Step 4.
Capability inventory: The agent has access to tool discovery, data retrieval from Gorgias, and the ability to send messages to Slack channels.
Sanitization: While no automated sanitization is implemented for the ticket content, the skill requires a manual review step ('Show the digest to the user for review') before posting to Slack. Furthermore, the disable-model-invocation: true setting in the frontmatter provides a platform-level restriction that prevents the model from autonomously invoking tools based on input data.

weekly-digest