The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted user input during the role discovery and requirement gathering phases (SKILL.md, steps 1 and 2). This data is interpolated into a markdown template and written to a local file using a filename derived from the subject (to-questionnaire-.md). Evidence chain: Ingestion points: User dialogue in conversational steps 1 and 2. Boundary markers: The skill uses Markdown headers to delimit sections but lacks explicit instructions to ignore potentially malicious embedded content. Capability: Writing local markdown files. Sanitization: The skill does not define specific escaping or validation for user input. This surface is considered acceptable as it is the primary intended function of the text-formatting tool.
[SAFE]: The skill configuration and instructions do not contain any evidence of credential harvesting, remote code execution, or persistence mechanisms. No obfuscated content or suspicious external dependencies were identified.

to-questionnaire