k2f-invitation
Pass
Audited by Gen Agent Trust Hub on Oct 1, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to install a core dependency using the command
npx skills add suzheng/k2f --skill k2f. This is a platform-specific setup step to incorporate the 'k2f' base engine authored by the same vendor. - [INDIRECT_PROMPT_INJECTION]: The skill identifies an attack surface by processing user-provided event facts and tone notes to generate invitation copy.
- Ingestion points: Event details and aesthetic requirements are ingested from user input during the generation of the
content_and_design.mdfile (SKILL.md). - Boundary markers: The instructions use structured sections like "Invitation lock" and "Message stack" to organize content, which helps separate data from instructions, although explicit escaping is not mentioned.
- Capability inventory: The skill is limited to generating semantic content and design intent for invitations; it does not perform sensitive file operations, network exfiltration, or unauthorized command execution.
- Sanitization: No specific sanitization or filtering of user input is documented, as the skill relies on the AI's adherence to the defined design hierarchy and rules.
Audit Metadata