The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the GitHub CLI (gh) and git to manage worktrees, issues, and pull requests throughout the audit cycle.
[REMOTE_CODE_EXECUTION]: The audit execution relies on external logic defined in amplihack recipes and DEFAULT_WORKFLOW.md, which are used during the analysis and fix phases.
[PROMPT_INJECTION]: The skill is subject to indirect prompt injection (Category 8) because it ingests data from an untrusted codebase at target_path. This information is then used to populate GitHub issues and PRs, potentially influencing the agent's behavior during automated remediation steps.
Ingestion points: Files within the codebase being audited.
Boundary markers: Documentation examples utilize shell heredocs (cat <<'EOF') to separate findings from command logic, which helps mitigate direct shell injection.
Capability inventory: Capabilities include file system manipulation, command execution, and network communication via the GitHub CLI.
Sanitization: No explicit sanitization or filtering of codebase content is described before it is interpolated into issue and pull request templates.
[EXTERNAL_DOWNLOADS]: The skill references external vendor-provided resources such as the quality-audit-cycle.yaml recipe and DEFAULT_WORKFLOW.md instructions.

quality-audit