The Agent Skills Directory

[SAFE]: The instructions focus on best practices for software engineering, such as code readability and modularity. No malicious intent, obfuscation, or data exfiltration patterns were detected in the skill logic.
[COMMAND_EXECUTION]: The skill uses the Bash tool to execute git diff for obtaining code changes. This is a legitimate use of the tool for a code review task.
[PROMPT_INJECTION]: The skill processes untrusted code changes from external inputs, which is a potential surface for indirect prompt injection. This is a characteristic of code analysis tools rather than a vulnerability in the skill itself.
Ingestion points: File paths provided via $ARGUMENTS and output from git diff as described in SKILL.md.
Boundary markers: Absent; the prompt does not provide specific delimiters to separate code from instructions.
Capability inventory: Tools for file system access (Read, Grep, Glob) and shell execution (Bash).
Sanitization: No input validation or sanitization is applied to the code being reviewed.

code-review