rspack-perf
Pass
Audited by Gen Agent Trust Hub on Sep 24, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute several shell commands, including
pnpm run build:cli:dev,pnpm run test:unit,cargo clippy, andgh workflow run. These commands execute local code or interact with the GitHub API. - [COMMAND_EXECUTION]: The GitHub CLI command
gh workflow run 'Ecosystem Benchmark' -f pr=<pr-number>uses a variable (<pr-number>) that is interpolated into the shell command. If this value is not strictly validated as an integer, it creates a surface for shell command injection. - [INDIRECT_PROMPT_INJECTION]: The skill follows a workflow where it ingests and modifies project code and subsequently executes build and test scripts on that modified code.
- Ingestion points: Target Rspack features or files identified for optimization (SKILL.md).
- Boundary markers: None identified; instructions do not specify delimiters for external code.
- Capability inventory: Execution of
pnpm,cargo, andghCLI tools (SKILL.md). - Sanitization: None identified; modified code is directly compiled and tested.
Audit Metadata