k2f-conference-handout
Pass
Audited by Gen Agent Trust Hub on Oct 1, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted user content such as talk outlines, slide outlines, and session notes to generate a handout blueprint. This creates a surface for indirect prompt injection attacks where malicious instructions hidden in the user content could potentially influence the agent's behavior.
- Ingestion points: The skill uses user-supplied talk outlines and session notes as the primary input for generating the
content_and_design.mdfile. - Boundary markers: Absent; there are no explicit delimiters or instructions to the agent to ignore potentially malicious embedded instructions in the source material.
- Capability inventory: The skill itself does not invoke high-risk capabilities like shell commands, network access, or dynamic code execution.
- Sanitization: Absent; the instructions do not mention any sanitization or validation of the input text.
- [EXTERNAL_DOWNLOADS]: The skill requires the installation of a secondary component from the same vendor to function correctly.
- Evidence: The installation command
npx skills add suzheng/k2f --skill k2fis listed in theSKILL.mdfile.
Audit Metadata