csdlc

Set up CSDLC for a project — generate WORKFLOW.md, create Foundry project space, scaffold design doc

Decompose a design doc into implementable stories (Step 1)

Generate a CSDLC Step 0 design doc for a feature — problem statement, scope, stories, open questions

Execute stories with sub-agents — work, review, wait for approval before merging

Hand off a story to a sub-agent (intern) for execution

Create a GitHub issue — bug, feature, tech debt, or task. Specify repo with --repo flag.

Package completed work for human review (Step 6) — async-friendly, iPad-optimized

Run QA against acceptance criteria (Step 5) — verify, test edge cases, provide evidence

Refine anything — a design doc, an idea, an issue, a plan. Challenge assumptions, flag gaps, resolve ambiguity.

End a CSDLC session — write handoff to Foundry NEXT.md

Structured review of agent output (Step 4) — scope check, tests, acceptance criteria

Run all stories in an epic autonomously — Hurricane Pattern. Full send.

Start a CSDLC session — load context, deliver standup. Optional persona arg (e.g., /standup clay)

Triage open GitHub issues — categorize, generate stories, hand off to /execute

Set up CSDLC for an existing project — generate WORKFLOW.md, create Foundry project space, scaffold a design doc. Use when the user says "bootstrap", "init", "set up CSDLC", "onboard this project", or when /standup detects no NEXT.md and no WORKFLOW.md.

Decompose a design doc into implementable stories (CSDLC Step 1). Read a design doc from Foundry or local file, break it into scoped stories with acceptance criteria. Use when the user says "breakdown", "stories", "decompose", "Step 1", or "break this into tickets".

Craft a sub-agent prompt from a refined CSDLC ticket, mirroring PROCESS.md Step 2 (Agent Prompt Crafting). Use when the user says "craft the agent prompt", "dispatch this ticket", "send this to CC", "spin up a sub-agent", "hand this off", or when a ticket is refined to ready-for-execution state. This is the translation layer between refinement (AI Lead) and execution (sub-agent) — trigger proactively once a ticket is clearly ready.

Generate a CSDLC Step 0 design doc for a feature, epic, or project. Includes bundled templates for three levels (project, epic, subsystem). Use when the user says "doc", "design doc", "Step 0", "write a design doc", or passes a feature description.

Execute stories with sub-agents — work the stories, review each PR, wait for human approval before merging. Bundles Steps 3 (execution), 4 (review), and optionally 5 (QA). The supervised execution mode. Use when the user says "execute", "work the stories", "run these", "go build this", or after /breakdown produces stories.

Hand off a single story to a sub-agent (intern) for execution. Reads the story from Foundry, crafts the agent prompt, spawns the intern in a worktree, writes results back to Foundry. Use when the user says "intern", "hand this off", "send it to the intern", "dispatch", "execute this story".

Create a GitHub issue from the current context — bug reports, feature requests, tech debt items. Use when the user says "file an issue", "open a bug", "track this", "create an issue", or when a problem is discovered during a session that needs tracking.

Package completed work for human review (CSDLC Step 6). Summarize what shipped, link PRs, flag judgment calls. Optimized for async review. Use when the user says "present", "show me what shipped", "package this", "ready for review", or after /qa passes.

The CSDLC methodology reference — pipeline steps, roles, refinement practices, rituals, context management. Use whenever you need to understand how CSDLC works, when a user asks about the process, when you're deciding which pipeline step applies, or when crafting agent prompts. Also triggers when terms like "refinement", "design doc", "Step 0", "Lightning Strike", "managed agent", "Hurricane Pattern" come up without clear context.

Run QA against acceptance criteria for a completed story (CSDLC Step 5). Verify each criterion, test edge cases, provide evidence. Separate from implementation — dedicated QA step. Use when the user says "qa", "test this", "verify", "does it work", or after /review passes.

Write the session handoff into NEXT.md at the end of a CSDLC working session. Use when the user says "wrap up", "session end", "retro", "update NEXT.md", "close out", "good stopping point", "let's pick this up tomorrow", or when a session is clearly ending. Trigger proactively — a missing handoff costs real time next session.

Structured review of sub-agent output (CSDLC Step 4). Run the AI Lead review checklist before presenting work to the human. Use when the user says "review", "check this", "did the intern do it right", or after a sub-agent completes a story.

Run all stories in an epic autonomously — the Hurricane Pattern. Read the dependency graph, dispatch stories to sub-agents in order, review PRs, merge, report back. Use when the user says "shipit", "run the whole epic", "full send", "hurricane", or "lightning strike this".

Run the CSDLC session-start ritual — load context, deliver the 5-part standup, activate the full CSDLC process. This is the entry point for the entire methodology. Use when the user says "standup", "catch up", "what are we working on", "bootload", "let's get started", or at any CSDLC session start. Accepts an optional persona argument (e.g., /standup clay).

Triage open GitHub issues — categorize as Step 1 ready or needs design doc, generate stories for actionable ones, hand off to /execute. Use when the user says "triage", "triage issues", "what can we fix", "check the issues", or wants to batch-process open bugs and features.