The Agent Skills Directory

[COMMAND_EXECUTION]: The skill enables the creation and interaction with isolated execution environments via the Sandbox SDK. This allows the agent to run LLM-generated code and shell commands for data analysis and computation (references/sandbox_sdk.md).
[EXTERNAL_DOWNLOADS]: The skill instructions direct the agent to fetch documentation from the official domain at https://docs.tensorlake.ai/llms.txt when bundled reference files are insufficient (SKILL.md).
[DATA_EXFILTRATION]: The skill handles authentication credentials for the TensorLake platform and third-party AI services. It promotes standard security practices such as using environment variables and secure secrets management within its application functions (SKILL.md, references/integrations.md).
[PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection because it ingests and processes data from external URLs and various document formats.
Ingestion points: External data is ingested through the fetch_page function and the DocumentAI.parse API (SKILL.md, references/documentai_sdk.md).
Boundary markers: No specific delimiters or instructions to ignore embedded commands are present in the documentation for handling fetched content.
Capability inventory: The agent has access to powerful tools including file system operations and command execution within the Sandbox environment (references/sandbox_sdk.md).
Sanitization: There is no evidence of sanitization or validation of untrusted content before it is processed by the agent.

tensorlake