The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to process and analyze text provided by users. This creates a potential surface for indirect prompt injection where malicious instructions could be embedded in the analyzed text. However, the skill provides specific structured output formats and analytical frameworks that guide the agent to treat the input as an object of linguistic study rather than a set of instructions. Furthermore, the skill lacks any high-risk capabilities such as network access, file system modifications, or command execution that could be leveraged in an attack.
[NO_CODE]: The skill package is composed exclusively of Markdown documentation and reference files (SKILL.md, references/cda-framework.md, and examples/sample_scenario.md). It contains no executable scripts (Python, JavaScript, Shell) or dependency manifests that could introduce security risks through code execution.

hum-discourse