The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill requires the installation of the typesafe-sdk package, which is an external dependency used to interact with the classification service.\n- [DATA_EXFILTRATION]: The skill transmits data provided in the state parameter to the typesafe.ai API for classification purposes.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted input through the Jev classifier, which is susceptible to adversarial content intended to influence classification results.\n
Ingestion points: The state parameter in the client.system_one call in SKILL.md accepts arbitrary text data.\n
Boundary markers: None defined in the skill instructions to separate untrusted data from instructions.\n
Capability inventory: The provided example script performs no dangerous actions (it only prints results), but downstream agent actions based on these results could be affected.\n
Sanitization: No sanitization or validation of the input text is described.

use-jev