Audit-Verlauf

The static findings are false positives caused by Markdown formatting, documented Git commands, dedicated worktree paths, and standard error redirection. However, the skill automatically runs repository-controlled setup and test commands, and it may commit a .gitignore change without explicit approval.

The 39 static findings are false positives caused by Markdown formatting, intended Git operations, documented worktree paths, and standard /dev/null redirection. Semantic review found automatic project code execution, an unapproved commit workflow, and a quoted-home-path defect.

Most static findings are false positives caused by Markdown fences, documented worktree paths, and stderr redirects. I found no prompt injection or data exfiltration intent. The main risks are automatic project command execution and an automatic commit step.

Static command execution findings are true positives because the skill instructs agents to run Git, package manager, build, and test commands. The hidden file and weak cryptography findings are false positives or low-risk context; no malicious intent, network exfiltration, credential access, or prompt injection was found.

Documentation-only skill providing git worktree management instructions. All 47 static findings are false positives: JSON metadata field names were misidentified as cryptographic patterns, bash code examples in documentation blocks were flagged as shell execution, and standard git directory patterns were marked as hidden file access. No executable code, network calls, or sensitive data exposure exists in this skill.

Documentation-only skill providing git worktree management instructions. All 47 static findings are false positives: JSON metadata field names were misidentified as cryptographic patterns, bash code examples in documentation blocks were flagged as shell execution, and standard git directory patterns were marked as hidden file access. No executable code, network calls, or sensitive data exposure exists in this skill.

Prompt-based skill with pure instructions. Uses standard git commands for worktree creation. No executable code, no network access, no sensitive data exposure. All capabilities align with stated purpose of creating isolated git workspaces.

Prompt-based skill with pure instructions. Uses standard git commands for worktree creation. No executable code, no network access, no sensitive data exposure. All capabilities align with stated purpose of creating isolated git workspaces.

Prompt-based skill with pure instructions. Uses standard git commands for worktree creation. No executable code, no network access, no sensitive data exposure. All capabilities align with stated purpose of creating isolated git workspaces.