open-code-review
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill instructs the agent to install the
@alibaba-group/open-code-reviewpackage from the NPM registry if it is not found on the system. This is an expected operation for this tool and targets a repository owned by the skill's author. - [COMMAND_EXECUTION]: The skill relies on executing the
ocrCLI tool to perform code reviews, check rules, and manage configuration settings. It also usesnpmfor installation. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data in the form of Git diffs and source code files. Because this content is sent to an LLM for review, there is a potential surface for indirect prompt injection if the code contains malicious instructions.
- Ingestion points: Reads Git diffs, staged/unstaged changes, and untracked files from the workspace.
- Boundary markers: The skill instructions do not specify any explicit delimiters or boundary markers for the code being reviewed; it relies on the internal handling of the
ocrtool. - Capability inventory: The skill can execute shell commands (
ocr,npm), write review results to files using the--outputflag, and modify its own configuration files. - Sanitization: No specific sanitization or filtering of the code content is described in the skill instructions before the data is processed.
Audit Metadata