k2f-sales-deck
Pass
Audited by Gen Agent Trust Hub on Oct 1, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and summarize external data like discovery notes, call transcripts, and CRM context to build sales presentations. This creates a surface for indirect prompt injection if the source data contains adversarial instructions.
- Ingestion points: External call notes, CRM context, and prospect-provided text (SKILL.md).
- Boundary markers: The skill uses specific Markdown blocks for content organization but does not define explicit delimiters or instructions to ignore embedded commands within the input data.
- Capability inventory: The skill generates structured content and interacts with the 'k2f' tool for compilation and export.
- Sanitization: Instructions include checks to ensure data integrity (no fabricated ROI), but no technical sanitization for injection is present.
- [EXTERNAL_DOWNLOADS]: The skill instructs the agent to install a secondary skill from the same vendor's repository.
- Evidence: Instruction to execute
npx skills add suzheng/k2f --skill k2ffound in SKILL.md. - Context: The resource originates from the same author (suzheng) as the skill being analyzed.
Audit Metadata