Skip to content

feat(ux-optimization): live-site audit + core-objective prompts (ADR 0014) - #35

Merged
chrispezza merged 1 commit into
mainfrom
feat/live-site-audit-prompts
Jul 19, 2026
Merged

feat(ux-optimization): live-site audit + core-objective prompts (ADR 0014)#35
chrispezza merged 1 commit into
mainfrom
feat/live-site-audit-prompts

Conversation

@chrispezza

Copy link
Copy Markdown
Collaborator

Summary

  • The full quill-router deliverable pass exposed two gaps in the just-shipped reverse pass: live-site auditing was deferred (ADR 0013) yet proved to be where code-suspected defects become measured findings — and where an entire render-only defect class lives (theme seams, orphans, missing responsive collapse) — and the strategy layer had no core_objective producer, leaving gap-analysis priorities anchored on an implicit objective.
  • ADR 0014 (Accepted, owner-directed) resolves both: 07_objective_metrics (Tier 1, always; gap analysis now requires core_objective) and 10_live_site_audit (Tier 2, context_gated on deployed URL + browser tooling; measure-don't-eyeball rules, both themes/viewports, DOM verification before reporting blank renders). Gap analysis and optimization spec renumbered to 08/09; prompt name: fields unchanged so checks and registry history are unaffected.

Type of change

  • Bug fix
  • New feature (non-breaking)
  • Breaking change
  • Prompt content change
  • Schema / compilation / validation change
  • Documentation only
  • Build / CI / tooling

Affected components

  • Plugin (plugin/) — commands, skills, subagent
  • Prompts (prompts/dev/)
  • Compilation pipeline (scripts/compile_spec.py)
  • Validation pipeline (scripts/validate_spec.py)
  • Handoff generator (scripts/generate_handoff.py)
  • MCP server (mcp/prompt-server/)
  • Documentation (docs/, README.md, CLAUDE.md)
  • CI (.github/workflows/)

Test plan

  • checks/run_checks.py green — zero findings on new/renamed files (frontmatter, placeholders, dependency graph, word limits, filename convention all cover them automatically)
  • Methodology field-tested before encoding: the live-site prompt's rules are transcribed from an actual browser audit of the pilot target (measurements, both themes, mobile, DOM-verification-of-blanks all exercised for real)
  • CI green on this PR

ADR impact

  • No ADR needed
  • New ADR added: docs/adrs/0014-live-site-audit-and-objective.md (Accepted — owner directed the capability; resolves ADR 0013's deferred item)
  • Existing ADR updated: —

Notes for reviewers

  • The renames (07→08, 09) keep filename order matching execution order; name: fields are stable so word_limits.json keys and dependency checks are untouched.
  • 10_live_site_audit is deliberately Tier 2: code-only runs stay valid, and headless environments lose nothing silently — the gate names both conditions.
  • Pilot evidence for every rule is in the pilot notes' new live-site addendum.

🤖 Generated with Claude Code

…R 0014)

Resolves ADR 0013's deferred live-site item and the missing objective
artifact, both surfaced by the full quill-router deliverable pass:
- 07_objective_metrics (core_objective; gap analysis now anchors on it)
- 10_live_site_audit (Tier 2, context_gated; measure-don't-eyeball rules,
  both-themes/both-viewports, DOM verification before reporting blanks)
- gap_analysis/optimization_spec renumbered 08/09 and wired to consume
  core_objective and live_site_audit
- SKILL.md + CLAUDE.md tables, word limits, pilot-notes addendum, v0.3.0

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
@chrispezza
chrispezza merged commit ca7be43 into main Jul 19, 2026
2 checks passed
@chrispezza
chrispezza deleted the feat/live-site-audit-prompts branch July 19, 2026 00:24
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant