optimize-bun-code
Pass
Audited by Gen Agent Trust Hub on Sep 12, 2026
Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The instructions direct the agent to run Bun CLI utilities including
bun --cpu-prof,bun --cpu-prof-md, andbun --inspect-brkto analyze application performance. - [DATA_EXFILTRATION]: The skill utilizes the
writeHeapSnapshotfunction to capture process memory for analysis. These snapshots can contain sensitive information like environment variables or session tokens that happen to be in memory at the time of the dump. - [DYNAMIC_EXECUTION]: The skill executes user-provided JavaScript and TypeScript files as part of the benchmarking and profiling workflow.
- [INDIRECT_PROMPT_INJECTION]: The skill operates by ingesting and executing benchmark scripts provided by the user, creating a surface where malicious instructions in those scripts could potentially influence the agent's behavior. • Ingestion points: The agent reads and executes user-provided scripts like
bench/workload.ts. • Boundary markers: Guidance is provided to the agent to validate request-derived paths before file access. • Capability inventory: The skill demonstrates capabilities for file system modification (Bun.write,mkdir) and process memory capture. • Sanitization: Best practices for input validation are recommended to prevent path traversal when handling external data.
Audit Metadata