The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from several external sources, including web scraping results (via the scrape run module) and database records. This content is integrated into the agent's context without explicit boundary markers or sanitization, creating a surface for indirect prompt injection attacks.
Ingestion points: External data is ingested via teable record get, teable sql-query, and the web scraping functionality described in guides/scrape-guide.md.
Boundary markers: The instructions do not prescribe the use of delimiters or warnings to prevent the AI from obeying instructions embedded in the retrieved content.
Capability inventory: The agent possesses high-impact capabilities, including full database CRUD, sending emails via the send-email command, and deploying automation scripts.
Sanitization: No sanitization or validation routines are defined for data retrieved from external sources before it is used to drive subsequent actions.
[DYNAMIC_EXECUTION]: The skill enables the generation and deployment of JavaScript code for server-side automations using the teable automation generate-script command. While this is a core functionality of the Teable platform, it allows for the runtime creation and execution of code based on AI-generated content.
[COMMAND_EXECUTION]: The skill relies extensively on the teable CLI for its operations. It includes a local installation script, scripts/install.sh, which automates the setup of the vendor's command-line tools and authentication.
[EXTERNAL_DOWNLOADS]: The provided installation script downloads and installs the official @teable/cli package from the public npm registry. These downloads are associated with the vendor's official infrastructure.

teable-assistant-ops