hackathon-deck-demo

Pass

Audited by Gen Agent Trust Hub on Oct 1, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill downloads and installs the playwright package and the @antasphere/slideless CLI tool to the user's cache directory and global environment. These are used for automated browser testing and deck publishing respectively.
  • [COMMAND_EXECUTION]: The skill executes multiple shell commands, including npm, npx, node, and vendor-specific tools like hackathon and slideless. These operations are necessary for setting up the environment, running tests, and publishing the final product.
  • [DYNAMIC_EXECUTION]: The skill involves two forms of dynamic execution. First, it generates and modifies a Playwright script (play.mjs) based on the application's structure to automate screenshot capture. Second, the fill.mjs script executes a render.mjs script provided by the Slideless template to generate the final HTML deck. These actions are central to the skill's primary purpose.
  • [INDIRECT_PROMPT_INJECTION]: The skill reads local repository files (README, routes, dashboard pages) to propose demo "episodes." This presents an indirect prompt injection surface where malicious content in the repository could influence the agent's proposed actions. However, the skill requires a manual confirmation step from the user before proceeding, mitigating the risk.
  • [DATA_EXFILTRATION]: The skill captures screenshots of the local application and pushes the resulting deck content to the Slideless platform via the slideless push and slideless share commands. This is the intended behavior for hosting the demo deck. The check.mjs script specifically monitors and blocks unauthorized external network requests during the deck verification phase.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 1, 2026, 04:52 PM