presentation-builder
Pass
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: SAFEREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [REMOTE_CODE_EXECUTION]: The skill downloads and executes code from a remote repository at runtime. In SKILL.md, it directs the agent to run bunx with a GitHub reference (github:neuromechanist/agentic-presentation-builder#v0.1.8) or to clone and install the repository. These are identified as vendor-owned resources based on the author context.
- [COMMAND_EXECUTION]: The skill executes shell commands via the apb CLI to validate, serve, and export presentations. These commands are run through bunx or bun run within the agent's environment.
- [INDIRECT_PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection. Ingestion points: User-provided topics, outlines, and slide content enter the agent context through the planning and authoring steps described in SKILL.md. Boundary markers: No delimiters or explicit instructions are provided to the agent to ignore potentially malicious instructions embedded in the user content. Capability inventory: The skill has the ability to execute shell commands via the apb CLI. Sanitization: No sanitization or validation of the user input is mentioned before it is incorporated into the presentation JSON.
- [EXTERNAL_DOWNLOADS]: The skill downloads its presentation engine from the author's GitHub repository.
Audit Metadata