| name | ks-engineer |
| description | Engineering — implementation, code review, architecture, and CI |
Route engineering requests to the appropriate DeepWork workflow.
Use this skill for implementation tasks, code review, architecture decisions,
bug fixes, refactoring, and CI/CD work.
Supporting references
Before starting implementation, review the relevant convention files
(co-located in this skill directory) for standards and process:
- Software engineer role: software-engineer.md -- implementation behavior, scope discipline, output format
- Code reviewer role: code-reviewer.md -- review focus, security checks, style enforcement
- Architect role: architect.md -- system design, trade-off analysis, spec writing
- Feature delivery: process.feature-delivery.md -- end-to-end lifecycle from issue through merged PR
- Pull requests: process.pull-request.md -- PR structure, squash merge rules, demo requirements
- Code review ownership: process.code-review-ownership.md -- review assignment and approval flow
- Continuous integration: process.continuous-integration.md -- CI pipeline standards
- Refactoring: process.refactor.md -- safe refactoring process
- Version control: process.version-control.md -- commit discipline, conventional commits
- Version control (advanced): process.version-control-advanced.md -- rebase, conflict resolution, lock files
- VCS context continuity: process.vcs-context-continuity.md -- PR progress tracking and resumability
- Git repos: process.git-repos.md -- repo cloning, worktree layout
- Shell scripts: code.shell-scripts.md -- shell script standards
- Code comments: code.comments.md -- commenting standards
- Nix devshell: tool.nix-devshell.md -- project-specific Nix environments
Available workflows
engineer
- engineer/implement -- implement a feature, fix, or refactor with TDD and quality gates
- engineer/doctor -- diagnose and fix engineering environment issues
platform_engineer
- platform_engineer/incident_investigation -- full incident investigation from triage to report
- platform_engineer/quick_investigate -- rapid incident triage without formal report
- platform_engineer/doctor -- debug local dev environment issues and unblock developer
- platform_engineer/error_tracking -- set up exception monitoring (Sentry, Honeybadger, etc.)
- platform_engineer/dashboard_management -- inspect and develop Grafana dashboards for services
- platform_engineer/cicd_optimization -- review CI/CD pipelines and suggest optimizations
- platform_engineer/release_builder -- set up CI release pipeline, release branches, and release notes
- platform_engineer/infrastructure_audit -- document and assess infrastructure setup against conventions
- platform_engineer/cloud_spend -- analyze cloud costs and identify waste
- platform_engineer/vulnerability_scan -- security vulnerability scanning and review
- platform_engineer/observability_setup -- set up Prometheus/Loki/VictoriaMetrics monitoring stack
- platform_engineer/infrastructure_migration -- plan and execute infrastructure migrations with validation gates
- platform_engineer/soc_audit -- assess SOC 2 readiness and produce compliance document
- platform_engineer/platform_issue -- create a platform engineering issue from current context
Routing rules
- Implementation tasks, feature work, bug fixes -->
engineer/implement
- Code review requests --> read
code-reviewer.md, then review directly
- Architecture or design questions --> read
architect.md, then answer directly or start engineer/implement for spec-driven work
- Engineering environment issues -->
engineer/doctor or platform_engineer/doctor
- Incidents, outages, or triage -->
platform_engineer/incident_investigation or platform_engineer/quick_investigate
- Monitoring, dashboards, observability -->
platform_engineer/dashboard_management or platform_engineer/observability_setup
- CI/CD, release pipelines -->
platform_engineer/cicd_optimization or platform_engineer/release_builder
- Infrastructure, cloud costs, migrations -->
platform_engineer/infrastructure_audit, platform_engineer/cloud_spend, or platform_engineer/infrastructure_migration
- Security, compliance -->
platform_engineer/vulnerability_scan or platform_engineer/soc_audit
- Filing a platform issue -->
platform_engineer/platform_issue
- Keystone module or convention changes --> use
/ks-system instead
- Filing general issues for discovered problems --> use
/ks-system instead
- If unclear, ask the user which workflow to run before starting
How to start a workflow
- Call
get_workflows to confirm available workflows.
- Call
start_workflow with job_name: "engineer" or job_name: "platform_engineer", workflow_name: <chosen>, and goal: "$ARGUMENTS".
- Follow the step instructions returned by the MCP server.