agentleFS
UwayxtUwayxt/agent-skills/.agents/skills/cross-functional-review/SKILL.md

cross-functional-review

Skill4 starsChanged 3 months agoMITView on GitHub

Review through business, UX, UI, and technical lenses before shipping. Use when final review, is this ready to ship, review business + UX + UI + tech, launch readiness, or cross-functional check.

The file

We split this skill into the rules it gives an agent

10 rules. About 0 are common, 10 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈10

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    No single lens catches everything.

    line 8
  2. Distinctive:

    Run four review passes. Each pass focuses on a different lens.

    line 10
  3. Distinctive:

    Check against upstream product thinking artifacts:

    line 14
  4. Distinctive:

    Do the modules implemented trace back to the business model (from business-model-reading)?

    line 19
  5. Distinctive:

    Check against UX artifacts:

    line 25
  6. Distinctive:

    Check accessibility-review criteria: contrast ratios, keyboard navigation, focus indicators, tap targets, semantic HTML, ARIA.

    line 30
  7. Distinctive:

    Check against design system artifacts:

    line 36
  8. Distinctive:

    Do micro-interactions follow the motion spec (from micro-interaction-motion-design)?

    line 41
  9. Distinctive:

    Check the implementation itself:

    line 48
  10. Distinctive:

    After all four passes, conflicts will surface.

    line 61

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Cross-Functional Review

No single lens catches everything. Business sees revenue; UX sees flow; UI sees polish; engineering sees performance. A cross-functional review forces all four lenses onto the same artifact before it ships. The goal is not consensus — it's completeness.

Run four review passes. Each pass focuses on a different lens. Findings from one pass may conflict with findings from another — that's expected. Document all findings, then resolve conflicts.

Pass 1: Business alignment

Check against upstream product thinking artifacts:

  • Does this feature/product serve the stated business goal (from product-discovery / product-strategy)?
  • Is it scoped to MVP (from mvp-scoping), or has scope crept?
  • Does the monetization model still make sense with this implementation (from business-model-thinking)?
  • Have all stakeholder requirements been addressed (from stakeholder-requirement-mapping)?
  • Do the modules implemented trace back to the business model (from business-model-reading)?

Verdict: Aligned / Misaligned / Partially aligned (list gaps)

Pass 2: UX flow

Check against UX artifacts:

  • Walk through every user flow (from user-flow-mapping). Does the implementation match the designed flow, including error and edge paths?
  • Are error states handled? Empty states? Loading states? Each must have a designed response, not a blank screen or a browser default.
  • Is the information architecture intact (from information-architecture)? Can users find what they need from the nav?
  • Apply usability-heuristics-check mentally: visibility of system status, user control, consistency, error prevention, recognition over recall, flexibility, minimalism, error recovery, help.
  • Check accessibility-review criteria: contrast ratios, keyboard navigation, focus indicators, tap targets, semantic HTML, ARIA.

Verdict: UX complete / UX has gaps (list missing states/flows)

Pass 3: UI consistency

Check against design system artifacts:

  • Does every component use design tokens (from design-tokens)? Grep for raw hex codes, pixel values, and hardcoded shadows — each one is a consistency violation.
  • Does typography follow the system (from typography-system)? Check font sizes, weights, line-heights against the type scale. Any off-scale values?
  • Apply visual-hierarchy-review: focal point, size hierarchy, color contrast, spacing rhythm, alignment, CTA hierarchy.
  • Are component states complete (from component-style-guide)? Hover, focus, active, disabled, loading, error — all present for every interactive component?
  • Do micro-interactions follow the motion spec (from micro-interaction-motion-design)? Correct durations, easing curves, reduced-motion support?
  • Is @media (prefers-reduced-motion: reduce) handled?

Verdict: UI consistent / UI has violations (list each with severity)

Pass 4: Technical quality

Check the implementation itself:

  • Code structure: Is the code organized logically? Components in the right directories? Clear separation between data, UI, and routing layers?
  • Tests: Are critical paths tested? At minimum: happy path, primary error path, permission boundary.
  • Performance: Large bundles? Excessive re-renders? Unoptimized images? N+1 queries? Measure, don't guess.
  • Accessibility implementation: Does the rendered output actually pass the criteria from accessibility-review? (Contrast tool, keyboard test, screen reader check)
  • Responsive behavior: Does the layout adapt correctly at every breakpoint defined in design-tokens?
  • Module sync: For dashboard projects, has module-registry-sync been run? Is the manifest valid? Does the dashboard shell show the module correctly?
  • Route integrity: Has route-integrity-checker been run? Are there dead-end CTAs or pending intents?

Verdict: Tech quality pass / Tech quality has issues (list with severity)

Pass 5: Conflict resolution

After all four passes, conflicts will surface. Handle them:

  • Business vs UX: Business wants a feature that UX says is confusing → document both perspectives, propose a compromise that serves the business goal without the UX pitfall. If no compromise exists, escalate with both sides clearly stated.
  • UI vs Engineering: UI polish requires engineering effort that delays launch → quantify the trade-off: "Completing the skeleton loading states adds ~4h of work but covers the 3rd most common user flow."
  • Scope vs Quality: Some findings are "nice to have" vs "must fix." Classify every finding: Blocker (ship is wrong without this), Warning (noticeable quality gap, not a showstopper), Note (improvement for next iteration).

Unresolved conflicts are escalated to the user with both sides clearly stated. Never ship with an unresolved blocker.

Completion criteria

Done when:

  • All 4 passes are complete with specific findings (not "looks fine")
  • Every finding is classified as blocker / warning / note
  • Every blocker has a resolution or is explicitly escalated
  • Conflict resolution pass has addressed cross-pass contradictions
  • Ship / no-ship recommendation is stated with rationale

Output format

Cross-functional review scorecard (Markdown artifact):

PassVerdictBlockersWarningsNotes
BusinessAligned/Misalignedcountcountcount
UX FlowComplete/Gapscountcountcount
UI ConsistencyConsistent/Violationscountcountcount
Technical QualityPass/Issuescountcountcount

Followed by:

  1. Detailed findings per pass (finding → severity → recommendation)
  2. Conflict resolution log
  3. Final recommendation: Ship ✅ / Don't ship ❌ / Ship with known issues ⚠️

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.