AI Agent Workflow Collection

Design stable APIs and interface contracts before implementation — REST/GraphQL endpoints, module boundaries, type contracts, FE/BE splits. Use when the user says "design the API", "contract this out", "split frontend and backend", or a slice crosses a boundary. Not for internal helpers or post-ship interfaces (use /rite-review).

Read-only audit dispatch for the active feature on the requested axis — security (OWASP, trust boundary, secrets), perf (measure-first, N+1, CWV), or simplify (Chesterton's Fence, deletion test). Use when the user says "security review", "is this safe", "is this fast enough", "perf check", "N+1", "simplify this", "Chesterton's Fence". Not for write actions or whole-project audits.

Prove UI behavior via the browser-proof ladder — browser-harness → Chrome DevTools MCP → /run+/verify → project E2E → manual — recording routes, viewports, screenshots, console, network, interactions, design-reference match to browser-evidence.md. Use when the user says "check the UI in browser", "screenshot this", "prove it renders". Not for backend-only features.

Debug systematically when tests, builds, or runtime/browser checks fail — deterministic feedback loop, reproduce, 3-5 ranked hypotheses, instrument, fix root cause, regression-test. Use when the user says "debug this", "why is it failing", "it broke", "the build is red", "the tests fail". Not for code review or flaky tests (fix the flake first).

Stress-test a single non-trivial decision via CLAIM → EXTRACT → DOUBT → RECONCILE → STOP with devrites-doubt-reviewer for an independent take. Use when the user says "are we sure", "double-check this", "what could go wrong", or a boundary / data-model / auth / public-API / migration change is about to commit. Not for post-merge review or trivial choices.

Implement frontend at senior designer-engineer quality — design-system discovery, shape-before-code, every state, a11y + WCAG 2.2 AA, anti-AI-slop, 2026 CWV bar, fullstack via contract-first slices. Use when the user says "build the UI", "ship-quality frontend", "design system", "a11y", or /rite-build detects UI. Not for polishing a built feature or design exploration.

One-question-at-a-time interview to extract what the user wants — each question carries a best-guess + structured options; stop at ~95% confidence. Use when the user says "interview me", "grill me", "I'm not sure what I want", or /rite-spec / /rite-define flags the ask as underspecified. Not for casual clarification or ideation (use rite-pressure-test).

Internal shared DevRites helper library. Documents cross-cutting engine commands and references; not a user workflow. Do not invoke directly.

Humanize prose without losing technical content. Use when asked to remove AI-writing tells, match a writing sample or project voice, edit artifacts/replies/commit or PR prose, or run /rite-polish's prose pass. Not for code comments or UI craft.

Refresh stale optional indexes. Use when structural lookup disagrees with live code or rite-doctor requests a synchronous reindex.

Verify uncertain framework or library behavior against official docs or installed source before relying on it, then record the source in evidence.md / decisions.md. Use when the user says "check the docs", "is this still the API", "don't guess", "cite the source", "verify this assumption", or build encounters an unfamiliar framework call. Not for project-internal code or trivial stdlib calls.

Plan the UX/UI before any code — produce the feature-level design-brief.md (design direction, key states, interaction model, visual-direction probe) the build then targets. Use when /rite-spec detects UI, or the user says "shape the UX", "plan the UI before coding", "design direction", "what should this look like", or a UI feature needs its look & flow decided up front. Woven into spec/build, not a separate phase. Not for building (use devrites-frontend-craft), polishing a built UI (use /rite-polish), or throwaway exploration (use /rite-prototype).

DevRites menu + router. Without args, renders the menu. With a verb (e.g. spec, build, review), dispatches to the matching rite- skill — /rite spec foo is equivalent to /rite-spec foo. Use when the user types /rite or names a phase verb.

Adopt an existing or legacy codebase into DevRites by reverse-engineering current behavior, seeding conventions, and establishing the workflow. Use for inherited or live applications.

Run the entire DevRites lifecycle end-to-end with no per-phase human iteration — spec → define → build×N → prove → polish → review → seal → ship — choosing the best option at every soft gate and recording the rationale. A vague prompt triggers an up-front clarifying interview; after that it runs unattended, pausing only on hard irreversible-risk gates (auth / migration / public-API / red tests), blocking or escalating gates, or a NO-GO. Default stops at the final type-GO; --ship (alias --yolo) auto-confirms it for a zero-touch push. Use when the user says "autocomplete", "do the whole thing", "run the full cycle", "build this end to end", "one-shot this feature", or "ship it autonomously". Not for a single phase (use the specific /rite- skill) or when the user wants to drive each step.

Implement exactly ONE vertical slice of the active feature, then stop with evidence — TDD, doubt loop, frontend craft, Spec Drift Guard. Use when the user says "build the next slice", "implement slice N", "continue", "code this slice". Not for bug fixes, prototypes, refactors outside scope, or two slices in a row.

Audit a completed spec for missing decisions before strategy or architecture. Use after /rite-spec when coverage is incomplete or stale; not for spec writing.

Converge intent and live code. Use when resuming a half-built feature, after /rite-adopt drift, or the user asks "what's left to build". Not for initial planning.

User-invoked helper for authoring DevRites overrides or extensions.

Decompose an approved spec.md into plan.md + vertical task slices + state.md; every acceptance criterion maps to ≥1 slice. Use when the user says "plan this", "break this into slices", "task breakdown". Not for writing code or repairing an existing plan (use /rite-plan).

Doctor DevRites setup or index health. Use when workflow wiring is broken, the user asks to check DevRites, or says "reindex". Not for application bugs.

Explicit browser dogfood QA for the active feature or branch.

User-invoked explainer that teaches one concept, diff, idea, or recent-work recap with an optional retrieval check-in.

Frame an ad-hoc ask before coding, then audit the diff. Use for underspecified imperative asks, raw-diff self-audits, or /rite-quick setup. Not a lifecycle gate.