The Agent Skills Directory

[COMMAND_EXECUTION]: The skill documents the use of the te-cli binary, allowing the agent to execute shell commands for data analysis and metadata management.
[CREDENTIALS_UNSAFE]: Mentions standard authentication mechanisms, including the TE_TOKEN environment variable and the token cache file located at ~/.te-cli/tokens.json.
[DATA_EXFILTRATION]: Describes network interaction with the ThinkingEngine backend (ta.thinkingdata.cn) and allows for user-defined host configurations.
[PROMPT_INJECTION]: Identified an indirect prompt injection surface where the agent processes external data returned by the CLI.
Ingestion points: Command output from te-cli domains (meta, analysis, audience, operation).
Boundary markers: Absent.
Capability inventory: Shell execution of te-cli and arbitrary HTTP requests via te-cli api.
Sanitization: None documented.

te-shared