pipecrew

Reads a list of repo paths and infers cross-repo topology from the code itself — OpenAPI specs, HTTP client calls, event publish/consume, DB connections, IAM/cross-stack references. Produces two Mermaid diagrams (high-level + detailed) WITHOUT requiring a pre-existing platform.md or config.json. Use when no workspace metadata exists, when a workspace was onboarded with an older plugin version, or when you want a code-grounded second opinion against an existing platform.md. Inputs the caller must provide: - repo_paths: list of absolute repo paths to scan - workspace_name (optional): label for the diagram title — defaults to the parent directory name of the first repo - output_mode: 'canonical' (emit both architecture-overview.mmd + architecture.mmd) | 'topic' (emit a single focused diagram) | 'audit' (read-only staleness report against existing diagrams) - topic (required when output_mode = topic): name of the focused concern (e.g., 'auth-flow', 'event-flow') - existing_diagrams (required when output_mode = audit): paths to existing .mmd files to compare against

Reviews AWS CDK (TypeScript / Python) stack implementations for INFRASTRUCTURE_IMPACT compliance, security defaults (encryption, IAM least-privilege, removalPolicy), cross-stack references, tagging, and test coverage. Produces a structured report with findings grouped by severity. Read-only — `cdk synth` / `cdk diff` output is consumed as a verification artifact alongside the source diff.

Implements new AWS CDK stacks in a TypeScript CDK project (S3 buckets, SQS queues, event notifications, IAM, CORS, Lambda functions, CloudFront, etc.). Reads the target repo's CLAUDE.md (and any context files it points to) plus existing stacks for naming conventions, stage/region handling, and resource patterns. Use for any TypeScript CDK repo. Inputs the caller must provide: - repo_path: absolute path to the CDK repo worktree - stack_name: the new stack's canonical name pattern (e.g., my-feature-{stage}{regionSuffix}) - requirements: what resources the stack must contain, with cross-references to other stacks/services - fix_list (optional): file:line targets with exact changes

Manages agent-facing context files (CLAUDE.md, agent-context/). Five modes: full (agent-context/ + CLAUDE.md, role-dispatched), claude-only (CLAUDE.md standalone), init (legacy — agent-context/ only), refresh (update after a feature ships), audit (staleness report, no writes).

Implements HTTP endpoints, views, serializers, models, migrations, and tests in a Django / Python service. Supports both API-first (OpenAPI spec present, typically via DRF + drf-spectacular) and code-first (no spec — architect inlines the endpoint contract) modes. Reads the target repo's CLAUDE.md for conventions, reads the existing code for patterns, then implements end-to-end. Inputs the caller must provide: - repo_path: absolute path to the target repo worktree - spec_policy: 'api-first' | 'code-first' - spec_file: relative path to the OpenAPI spec (only when spec_policy=api-first) - inline_contract: for code-first, the endpoint contract (method, path, request/response shape, status codes) extracted from the architect's API_DESIGN - feature_summary: one paragraph - requirements: FR/EC list - endpoints_to_implement: list of endpoint paths + methods - fix_list (optional): file:line targets for fix rounds

Standalone cross-repo assessment. Checks a feature branch across all repos for cross-repo integration: wire-shape agreement, requirement enforcement symmetry, event/infra wiring. Use for verifying manually-implemented features or pre-merge validation.

Audit or refresh PipeCrew context docs at three scopes: a single repo (its agent-context/, CLAUDE.md, and DESIGN_SYSTEM.md if frontend), the workspace (platform.md), or everything (workspace + every repo). Audit mode reports staleness only; refresh mode updates docs to match current code.

PipeCrew delivers a feature end-to-end across all workspace repos. Orchestrates: requirements → architecture → spec editing → parallel implementation (backend + frontend + mock + infra) → cross-repo assessment. Usage: /deliver <feature description>. Reads workspace config from {workspace_root}/{workspace}/config.json.

Discover a new project for the PipeCrew. Inspects repos, interrogates the domain, generates workspace config + platform.md + CLAUDE.md files + domain-specific agents. Run once per project, then use /deliver to ship features.

Generate or refresh architecture diagrams for a workspace. Default mode regenerates the two canonical Mermaid files (architecture-overview.mmd + architecture.mmd) for an onboarded workspace by re-dispatching the solution-architect in discovery mode. Optional --topic flag produces a focused diagram on a specific concern (auth flow, event flow, deploy topology, etc.) without touching the canonical files. Reuses the same conventions documented in rules/discovery-diagrams.md that /discover uses. Standalone — does not run the rest of /discover.