Engineering

Expert technical consultation — architecture questions, trade-off analysis, best practices.

Explore solutions with trade-off analysis. Ideation, architecture decisions, feasibility assessment.

Adversarial code review — security holes, false assumptions, failure modes. Supports PR number, commit hash, or codebase scan.

Set coding experience level for tailored explanations and output format.

Systematic debugging with root cause analysis. Bugs, test failures, performance issues, CI/CD failures, log analysis.

Deploy applications to various platforms with environment-appropriate strategies.

Extract structured knowledge from journals into categorized markdown files.

Search library/framework documentation via llms.txt (context7.com). API docs, GitHub repository analysis, latest features.

Discover available skills matching a task description.

Intelligent bug fixing — root cause analysis before applying targeted fixes.

Git operations with conventional commits. Staging, committing, pushing, PRs, merges. Auto-splits commits by type/scope. Security scans for secrets.

Smart feature implementation — plan, code, test, review cycle.

Capture session insights as episodic memory — aha moments, gotchas, design decisions, debugging breakthroughs.

Extract, compare, port, or adapt features from GitHub repositories into the current project.

Add OKF YAML frontmatter to project documentation so AI agents can grep by types and tags.

Create detailed implementation plans with phases, tasks, and architecture decisions.

Organize files, directories, and content structure. Determine output paths, standardize project layout.

Fast codebase scouting with parallel agents. File discovery, task context gathering, quick searches.

Security audit — OWASP top 10, auth vulnerabilities, dependency scanning, secrets detection.

Automated security scanning for common vulnerabilities and exposed secrets.

Step-by-step reasoning for complex multi-layer problems before implementation.

Unified ship pipeline — implement, simplify, test, review, commit in one command.

Create new Claude Code skills from description or existing pattern.

Transform brainstorm output into formal specifications with FR/SC/NF vocabulary and manual test generation.

Comprehensive testing — unit, integration, coverage analysis, error handling, performance validation.

Launch the manual test runner for a spec module. Auto-invoked by the implementation skill when an SC-M (manual-test) queue exists.

Generate LinkedIn technical articles from journals or notes. Short post (300-600w) and long article (800-1500w).