Command Development
Pass
Audited by Gen Agent Trust Hub on Oct 23, 2026
Risk Level: SAFECOMMAND_EXECUTIONDYNAMIC_CONTEXT_INJECTION
Full Analysis
- Instructional Command Execution Patterns: The skill documentation provides extensive guidance and examples for using the
Bashtool to interact with development utilities such asgit,npm,gh,kubectl, andhelm. These patterns are presented as best practices for building context-aware developer tools, with recommendations to use restrictive command filters likeBash(git:*)to limit the scope of execution. - Dynamic Context Injection Support: The skill details the usage of backtick syntax (
!command) to dynamically include shell command output in the agent's context. This is a documented platform feature for slash commands, and the skill provides appropriate examples for gathering repository state, such as `!`git statusand!git branch --show-current``. - Educational Security Testing: In the testing strategies reference, the skill includes a section on verifying that unauthorized commands are correctly blocked by the agent's safety layers. It uses a snippet of
!ls -la /`` specifically as a negative test case to ensure the platform's tool-use permissions are properly enforced, representing a security-conscious design pattern. - Input Validation Patterns: The skill explicitly teaches validation techniques to prevent command injection, such as using
grepwith strictly defined regex patterns (e.g.,grep -E "^(dev|staging|prod)$") to sanitize user-provided arguments before they are used in shell commands. - Standard Development Tool Integration: All external references, package managers (
npm,pip), and cloud service CLIs (aws,kubectl,gh) mentioned in the examples are well-known technology services and are used within their standard development use cases.
Audit Metadata