Getting Started¶
Prerequisites¶
- Agent Skills–compatible environment (e.g. Cursor IDE)
- Jira or Azure DevOps – Active account with API access
- GitHub – Account with personal access token
Step 1: Configure MCP¶
MCP combinations: - GitHub + Jira — GitHub for repos/PRs, Jira for issue tracking - ADO only — Azure DevOps provides both repo management and issue tracking
Full MCP setup → (Option A or B).
Open your environment’s MCP settings (e.g. Cursor: Settings → Features → Model Context Protocol) and add:
{
"mcpServers": {
"github": {
"url": "https://api.githubcopilot.com/mcp/",
"headers": {
"Authorization": "Bearer <TOKEN>"
}
},
"atlassian": {
"command": "npx",
"args": ["-y", "mcp-remote", "https://mcp.atlassian.com/v1/mcp"]
}
}
}
Get GitHub Token:
- Go to github.com/settings/tokens
- Generate new token (classic)
- Select scopes:
repoandread:org - Copy token into config above
Authorize Atlassian:
First time you use Jira, your environment may prompt for OAuth (e.g. Cursor prompts for Atlassian authorization).
Step 2: Install Skills¶
SDLC Workflow Skills uses the Agent Skills format. The same skills/ layout and SKILL.md files work for Cursor, Claude, and Codex—each environment loads from its own skills directory (see table below). We recommend installing with the npx skills CLI; Cursor users can alternatively install from GitHub or copy from a release.
| Environment | Project-level | User-level (global) |
|---|---|---|
| Cursor | .cursor/skills/ |
~/.cursor/skills/ |
| Claude | .claude/skills/ |
~/.claude/skills/ |
| Codex | .codex/skills/ |
~/.codex/skills/ |
Choose one of the following.
Option A (recommended): Install with npx skills¶
The skills CLI installs skills for Cursor, Claude, Codex, and other agents. Discover skills at skills.sh.
Cursor:
When the CLI asks Symlink or Copy, choose Copy. Cursor does not reliably list or load skills when ~/.cursor/skills is a symlink; using Copy avoids this.
For other agents or global install, see vercel-labs/skills and skills.sh.
Layout (Cursor, Claude, Codex)
The same layout works for all three: a top-level skills/ directory; each skill is a subfolder containing SKILL.md with YAML frontmatter name (matching the folder name) and description. See Cursor: Agent Skills, agentskills.io, and skills.sh.
To verify the install: open your agent chat and type /—you should see the SDLC commands (e.g. /create-task, /start-task). Or run /mcp-status to confirm MCP connections.
Option B: Download release and copy¶
Skills live in skills/ (in the repo or in the release archive). Copy them into your environment’s skills directory.
Summary:
- Download the latest release from the Releases page
- Extract the archive
- Copy
skills/*to your environment’s skills directory (see table above: Cursor.cursor/skills/or~/.cursor/skills/, Claude.claude/skills/or~/.claude/skills/, Codex.codex/skills/or~/.codex/skills/). Cursor Team/Enterprise: Use Dashboard → Team Content → Skills to create team skills fromskills/; they sync to your team. - Restart your IDE or agent
Step 3: Run Your First Skill¶
Open your agent chat (e.g. Cursor: Cmd/Ctrl + L) and run:
AI will:
- Read the command instruction
- Connect to Jira via MCP
- Create story with title, acceptance criteria, and labels
- Return story ID (e.g.,
PROJ-123)
That's it! You're using agentic skills.
Try a Development Workflow¶
How It Works¶
AI agents handle routine tasks:
- Story and epic creation
- Implementation planning
- Branch and PR creation
- Test generation
- Issue updates
You control decisions:
- Plan approval
- Code review
- Work selection
- Architecture
Full transparency:
- Specs in
specs/folder (permanent feature contracts) - Plans in
.plans/folder (transient task implementation) - PRs link to stories
- Commits include story IDs
See the product flow diagram →
Next Steps¶
- View All Skills - Browse available skills
- Quick Reference - Skill cheat sheet
References¶
- Cursor: Agent Skills - Skill format, GitHub install, Cursor/Claude/Codex skill directories
- Claude: Extend with skills - Claude Code skills (
.claude/skills/) - agentskills.io - Agent Skills open standard
- skills.sh - Agent Skills directory; vercel-labs/skills - npx skills CLI