SD0
Community@sd0xdev · Taipei, Taiwan
AI-native engineering leader (AI-as-default) since 2025—spec-to-code, automated reviews, and test generation with measurable quality gates.
Agent Skills by SD0
Showing 100 vetted skills indexed across 2 GitHub repositories.
git-profile
Manage git identity and GPG signing profiles across repositories.
op-session
Initialize and manage 1Password CLI sessions for non-TTY subprocess environments.
skill-health-check
Validate skill quality against routing, progressive loading, and verification criteria.
adr
Writes numbered Architecture Decision Records with superseded linking into feature documentation directories.
necessity-audit
Audits lifecycle specs for over-designed elements using a 6-dimension YAGNI rubric and Codex adversarial debate.
update-docs
Synchronizes documentation with current code state by researching changes and updating current-authority docs.
review-spec
Reviews technical specification documents for completeness, feasibility, risk, and code consistency.
ask
Answers codebase questions with intent classification and source-attributed context gathering.
contract-decode
Decodes EVM function selectors, calldata, and revert data using local cast tools and external ABI APIs.
verify
Runs lint, typecheck, unit, integration, and e2e verification steps across detected project ecosystems.
ui-first-principles
Generates JTBD-driven UI field-priority decisions and gap reports from API field sets.
codex-setup
Initializes AGENTS.md and installs git hooks for Codex CLI projects.
generate-runner
Generate a customized precommit runner script for Node, Python, Rust, or Go projects.
doc-refactor
Refactor documents by simplifying content and converting flows into sequenceDiagrams.
check-coverage
Assess unit, integration, and E2E test coverage and identify gaps with prioritized recommendations.
project-audit
Audit project health with 12 deterministic checks across five scored dimensions.
merge-prep
Analyzes Git branches before merging to detect conflicts and generate suggested merge commands.
deep-analyze
Analyzes proposals against codebase patterns and produces actionable implementation roadmaps with alternatives.
pr-summary
Lists open GitHub PRs, filters automation bots, and groups them by ticket ID into Markdown.
statusline-config
Generates and customizes the Claude Code statusline shell script with segments and color themes.
sharingan
Replicate skills from GitHub repos, articles, or local code into sd0x-dev-flow skill definitions.
deep-research
Orchestrates parallel research agents to synthesize web, codebase, and community evidence into scored reports.
recap-doc
Generates post-development recap documents from git scope reports with blind-spot detection.
project-brief
Convert technical specifications into PM/CTO-readable executive summaries.
Frequently Asked Questions About SD0
FAQPage SchemaWhat tasks can I accomplish with SD0's skill collection?▼
You can run the full development lifecycle: requirements analysis, tech specs, architecture docs, feature implementation, bug fixes, Codex second-opinion code reviews, OWASP security audits, test coverage gap analysis, smart commits, stacked PR merges, and CI monitoring — all as Claude Code skills.
Who is the target user for these skills?▼
Software engineers and engineering leads practicing AI-native development with Claude Code and Codex CLI. It suits developers who want enforced quality gates — precommit checks, pre-PR confidence audits, adversarial plan reviews — rather than unreviewed generated code.
How do I install and run these skills in a project?▼
Use the project-setup skill to auto-detect your Node.js/TypeScript framework and populate CLAUDE.md, or codex-setup to generate AGENTS.md and commit-msg hooks. install-rules, install-hooks, and install-scripts persist plugin assets into .claude/ for use without the plugin loaded.
What prerequisites and dependencies are required?▼
Skills require Claude Code, and many review skills depend on Codex CLI or Codex MCP. Git and the GitHub gh CLI are needed for PR and CI skills; Jira skills use Atlassian MCP; op-session needs the 1Password CLI; package-manager commands cover npm, pnpm, yarn, pytest, cargo, and go.
Do the skills support verification and quality gating before merging?▼
Yes. The verify skill runs lint, typecheck, unit, integration, and e2e stages; pre-pr-audit produces a 5-dimension readiness score; risk-assess detects breaking changes and blast radius; seek-verdict obtains independent Codex confirmation or dismissal of review findings.