supercharge
Pass
Audited by Gen Agent Trust Hub on Oct 1, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes various testing and linting commands to verify codebase integrity before and after refactoring. Commands include
npm test,pytest,go test,npm run lint,ruff,golangci-lint,tsc --noEmit, andmypy. These are standard developer tools triggered based on the auto-detected project stack. - [INDIRECT_PROMPT_INJECTION]: The skill represents a surface for indirect prompt injection because it ingests and processes entire codebase files for smell detection and refactoring. A malicious project file could potentially attempt to influence the agent's behavior.
- Ingestion points: Codebase files being scanned and refactored (SKILL.md, Phase 2).
- Boundary markers: The skill instructs the agent to treat sub-agent reports as claims and verify them against evidence (SKILL.md, Phase 2). It also uses specific, scoped prompts for scanning.
- Capability inventory: The agent can write to files (refactoring) and execute shell commands (tests/linters).
- Sanitization: The skill relies on 'Safe Refactoring Rules' and human approval flows to mitigate risks of unauthorized or dangerous changes.
Audit Metadata