audit-bundle-size
Pass
Audited by Gen Agent Trust Hub on Nov 5, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill uses
firecrawl:firecrawl_searchto research alternatives for large dependencies. This process ingests untrusted data from the web into the agent's context. Without explicit sanitization or boundary markers, malicious search results could potentially influence the agent to perform unsafe actions during the optimization process. - Ingestion points: External web search results in Phase 3.
- Boundary markers: Absent; instructions do not specify how to isolate search content from agent instructions.
- Capability inventory: Shell command execution via
npm,npx, and standard Unix utilities (find,xargs,ls), file system modifications, and network access. - Sanitization: No explicit validation or filtering of content retrieved from web searches.
- [COMMAND_EXECUTION]: The skill requires the agent to execute various shell commands for building the project, installing analysis tools, and verifying the application state. These include commands like
npm install,npm run build, andnpx playwright, which are powerful capabilities necessary for the skill's primary function but represent a standard security surface area.
Audit Metadata