code-refactoring-refactor-clean
Pass
Audited by Gen Agent Trust Hub on Apr 13, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection attacks.
- Ingestion points: User-provided code is accepted via the
$ARGUMENTSvariable inSKILL.mdand processed directly. - Boundary markers: No delimiters (e.g., XML tags, triple backticks) are used to separate instructions from the code being analyzed.
- Capability inventory: The skill instructions explicitly direct the agent to 'Apply changes' (file writes) and 'Update tests and verify regressions' (command execution).
- Sanitization: There is no evidence of validation or filtering of input data to prevent embedded instructions from being interpreted by the LLM.
- [COMMAND_EXECUTION]: The skill requires the ability to execute shell commands to fulfill its stated purpose.
- Evidence: The instructions in
SKILL.mdstate 'Update tests and verify regressions', which typically involves running test suites or build tools. - Context: While expected for a developer tool, this capability increases the impact of the indirect prompt injection vulnerability.
Audit Metadata