What You Get After Installation

Installing the CLI gives you a command-line workflow for human-in-the-loop agentic software development.

Core Capabilities

• Create structured runs tied to repository + task definitions.
• Create sprint plans with explicit goals and ownership.
• Track run status and step progression.
• Approve critical gates before execution advances.
• Create orchestrator handoffs to specialist agent roles.
• Record handoff completion back into auditable logs.
• Preflight-check MCP server availability for planned capability work.
• Fallback to local offline documentation capture via existing scraper scripts.
• Generate artifacts for planning, patching, testing, review, and PR drafting.
• Maintain auditable run history and timeline events.

Artifacts You Receive

• PLAN.md: scoped implementation plan
• PATCH.diff: proposed code diff
• TEST_REPORT.md: validation summary
• REVIEW.md: quality/security review notes
• PR_DRAFT.md: merge-ready PR draft text
• artifacts/sprints/<SPRINT_ID>/SPRINT-PLAN.md: sprint goals + ownership
• C:\code\_agent-instructions\HANDOFF.md: handoff entries from orchestrator
• C:\code\_agent-instructions\TASK-COMPLETE.md: completion notes from agents
• C:\code\_agent-instructions\offline_reference\...: local captured documentation for offline-first workflows

Human-in-the-Loop by Default

Runs pause at approval gates (plan, apply_patch, pr) so your team can review before moving forward.

Agent handoffs are auditable end-to-end: orchestrator messages, role completions, sprint artifacts, and capability preflight checks are all captured for traceability.