mike/kickstart.nvim
1
1
Fork 0
forked from Neptune/mirror.kickstart.nvim
Code Issues Pull requests Releases 40 Wiki Activity Actions

Release v1.27.0 — Writing Quality & Frontend Design Skills #52

Merged
Mike Bros merged 16 commits from release/1.27.0 into master 2026-03-16 04:26:36 +00:00
Conversation 3 Commits 16 Files changed 15 +1472 -231
Gravity Bot commented 2026-03-16 04:07:42 +00:00
Collaborator
Copy link

Release v1.27.0 — Writing Quality & Frontend Design Skills

Changes

Anti-Slop Writing Quality Skill (Epic OP#1930)

  • OP#1980: Create gt-writing-quality SKILL.md with 63 banned phrases, structural tic rules, and 5-dimension scoring rubric
  • OP#1982: Create before/after examples demonstrating improvement across wiki, journal, docs, and PR prose
  • OP#1984: Update gt-wiki and gt-journal to reference gt-writing-quality
  • OP#1985: Remove old Writing/ skill directory
  • OP#1987: Validate enforcement against sample prose

Frontend Design Quality Skills (Epic OP#1931)

  • OP#1988: Create domain/frontend/taste-design SKILL.md with tunable parameters and engineering constraints
  • OP#1989: Create domain/frontend/aesthetic-direction SKILL.md with creative design philosophy
  • OP#1990: Add taste parameter support to gt-project-setup and .gravity.yaml schema
  • OP#1991: Update domain/README.md with frontend design skills and usage guidance
  • OP#1992: Validate frontend skill structure and .gravity.yaml integration

Output Completeness Enforcement

  • OP#1932: Create domain/methodology/output-completeness skill

Infrastructure

  • OP#1971: Initialize version branch
  • OP#2193: Pre-release merge — consolidated parallel feature branches (1 trivial conflict in manifest.json resolved)

Checklist

  • All version tasks closed in Gravity PM
  • Epics #1930, #1931 in testing
  • Version file matches Gravity PM version (manifest.json → 1.27.0)
  • No credentials in tracked files

References

Version: v1.27.0 — Writing Quality & Frontend Design Skills (Gravity PM)

## Release v1.27.0 — Writing Quality & Frontend Design Skills ### Changes **Anti-Slop Writing Quality Skill** (Epic OP#1930) - OP#1980: Create gt-writing-quality SKILL.md with 63 banned phrases, structural tic rules, and 5-dimension scoring rubric - OP#1982: Create before/after examples demonstrating improvement across wiki, journal, docs, and PR prose - OP#1984: Update gt-wiki and gt-journal to reference gt-writing-quality - OP#1985: Remove old Writing/ skill directory - OP#1987: Validate enforcement against sample prose **Frontend Design Quality Skills** (Epic OP#1931) - OP#1988: Create domain/frontend/taste-design SKILL.md with tunable parameters and engineering constraints - OP#1989: Create domain/frontend/aesthetic-direction SKILL.md with creative design philosophy - OP#1990: Add taste parameter support to gt-project-setup and .gravity.yaml schema - OP#1991: Update domain/README.md with frontend design skills and usage guidance - OP#1992: Validate frontend skill structure and .gravity.yaml integration **Output Completeness Enforcement** - OP#1932: Create domain/methodology/output-completeness skill **Infrastructure** - OP#1971: Initialize version branch - OP#2193: Pre-release merge — consolidated parallel feature branches (1 trivial conflict in manifest.json resolved) ### Checklist - [x] All version tasks closed in Gravity PM - [x] Epics #1930, #1931 in testing - [x] Version file matches Gravity PM version (manifest.json → 1.27.0) - [x] No credentials in tracked files ### References Version: v1.27.0 — Writing Quality & Frontend Design Skills (Gravity PM)
Add comprehensive writing quality skill based on stop-slop methodology:
- Banned AI-tell phrase list (throat-clearing, emphasis crutches, jargon,
  filler adverbs, meta-commentary, performative emphasis, intensifiers)
- Banned structural tics (binary contrasts, dramatic fragmentation,
  rhetorical Q&A, formulaic constructions)
- Banned rhythm patterns (metronomic sentences, tripling, punchy endings,
  staccato fragmentation, em dash reveals, absolute words)
- Em dash rewrite rules with preference hierarchy (split > reword > colon)
- 5-dimension scoring rubric: Directness, Rhythm, Trust, Authenticity,
  Density (1-10 each, target 35+/50)
- Content Hygiene section merged from Writing/SKILL.md (footer normalization)
- Humanize workflow merged from Writing/Workflows/humanize.md (em dash rules,
  rewrite strategy, process)
- Trigger conditions and integration points
- Manifest entry for deployment

Source: https://github.com/hardikpandya/stop-slop

Closes OP#1980

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Adapted from the taste-skill repo (github.com/Leonxlnx/taste-skill).
Provides tunable DESIGN_VARIANCE, MOTION_INTENSITY, and VISUAL_DENSITY
parameters (1-10, default 5) with engineering constraints for layout,
typography, icons, color, motion, and content anti-slop rules.
Configurable per project via .gravity.yaml taste: section.

Closes OP#1988

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Create 5 before/after pairs demonstrating anti-slop rewriting:
- Wiki intro (16/50 -> 42/50)
- Journal entry (14/50 -> 41/50)
- Technical documentation (13/50 -> 44/50)
- PR summary (12/50 -> 45/50)
- Release notes (11/50 -> 44/50)

Each pair includes full 5-dimension rubric scoring (Directness, Rhythm,
Trust, Authenticity, Density) with specific notes explaining the scores.
Add manifest.json entry for the examples reference file.

Closes OP#1982

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- gt-wiki SKILL.md: update Create Page step 6 reference
- gt-wiki SKILL.md: update Integration Points table entry
- gt-wiki SKILL.md: update Quality Checks checklist item
- gravity-skills.tsv: rename Writing entry to gt-writing-quality
- gt-journal does not reference Writing (no changes needed)
- No other skills reference the old Writing/ path

Closes OP#1984

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Delete configs/claude/skills/Writing/SKILL.md
- Delete configs/claude/skills/Writing/Workflows/humanize.md
- Remove skill-writing and skill-writing-humanize manifest entries
- Add ~/.claude/skills/Writing/ to removed_configs for cleanup
- Verified no broken references remain in any skill file

Closes OP#1985

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Provides creative direction guidance adapted from Anthropic's
frontend-design skill philosophy. Covers aesthetic commitment (pick an
extreme and commit), typography/color/motion/layout philosophy with
direction examples, and clear delineation from taste-design (this skill
= what/why, taste-design = how/avoid).

Closes OP#1989

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Test results (all 5 test cases pass):

TC1 - Banned phrase list: PASS
  63 banned phrases/words across 9 categories (throat-clearing, emphasis
  crutches, business jargon, filler adverbs, meta-commentary, performative
  emphasis, telling-not-showing, AI intensifiers, formulaic transitions)

TC2 - Structural tic rules: PASS
  17 patterns across 4 categories (binary contrasts, dramatic fragmentation,
  rhetorical Q&A, formulaic constructions) plus 7 banned rhythm patterns

TC3 - Scoring rubric: PASS
  All 5 dimensions (Directness, Rhythm, Trust, Authenticity, Density) with
  1-10 scale, 4 score bands each with specific criteria, 35/50 threshold

TC4 - Humanize workflow merged: PASS
  Em dash rules (5 usage patterns, 3-tier rewrite preference), content
  hygiene (footer normalization rules), process steps all preserved from
  Writing/Workflows/humanize.md with no regression

TC5 - gt-wiki and gt-journal references: PASS
  gt-wiki SKILL.md updated in 3 locations (Create Page step 6, Integration
  Points table, Quality Checks checklist). gt-journal had no Writing refs.
  gravity-skills.tsv updated. No broken references remain.

Closes OP#1987

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Adds Step 5.5 to the setup wizard for frontend taste configuration
(DESIGN_VARIANCE, MOTION_INTENSITY, VISUAL_DENSITY dials 1-10).
Documents the taste: section in schema.md with field reference,
validation rules, example config, and behavioral descriptions
per parameter level.

Closes OP#1990

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Adds taste-design and aesthetic-direction to the frontend skill count
(9 -> 11) and documents their complementary roles: aesthetic-direction
for creative "what/why", taste-design for engineering "how/avoid".
Updates source attribution to include non-Antigravity skill sources.

Closes OP#1991

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Test results — all 5 test cases pass:

TC1: taste-design SKILL.md has valid frontmatter (name, description,
  source, risk: safe) and documents all 3 tunable parameters
  (DESIGN_VARIANCE, MOTION_INTENSITY, VISUAL_DENSITY) — PASS

TC2: aesthetic-direction SKILL.md has valid frontmatter (name,
  description, source, risk: safe) — PASS

TC3: Parameter defaults documented as 5/5/5 in both taste-design
  SKILL.md ("all parameters default to 5") and schema.md ("each
  parameter defaults to 5 (balanced)") — PASS

TC4: Custom parameter example present in .gravity.yaml schema with
  3 examples: full schema (5/5/5), field reference example (7/5/6),
  and npm project example (8/6/4) — PASS

TC5: gt-project-setup SKILL.md mentions taste configuration in 6
  locations including Step 5.5 wizard flow, parameter questions,
  YAML output example, and taste-design skill cross-reference — PASS

Closes OP#1992

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Brings in the anti-slop writing quality skill chain:
- gt-writing-quality SKILL.md with banned phrases, structural tics, scoring rubric
- Before/after examples reference document
- Updated gt-wiki and gt-journal refs
- Removed old Writing/ skill directory
- Validated enforcement against sample prose

Refs OP#2193

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Brings in the frontend taste/aesthetic design skill chain:
- taste-design SKILL.md with tunable parameters and engineering constraints
- aesthetic-direction SKILL.md with creative design philosophy
- gt-project-setup taste parameter support and schema updates
- domain/README.md updated with frontend design skills
- Validated integration and .gravity.yaml support

Conflict resolved: manifest.json — combined both branches' new entries,
dropped old skill-writing entry (already removed by Branch A)

Refs OP#2193

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Create domain/methodology/output-completeness/SKILL.md that prevents
lazy, truncated, or placeholder code output across all implementation
work. Bans placeholder comments, truncated code blocks, elided sections,
lazy variable names, and incomplete branches/error paths.

- Added to gr-implementation domain skill references (loaded for all work)
- Updated domain/README.md methodology count (5 → 6)
- Added manifest.json deployment entry

Closes OP#1932

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
chore(release): bump version to 1.27.0
All checks were successful
CI / lua-check (pull_request) Successful in 10s
Details
PR Validation / validate-branch (pull_request) Successful in 1s
Details
CI / security (pull_request) Successful in 14s
Details
PR Validation / validate-release-pr (pull_request) Successful in 6s
Details
CI / sast (pull_request) Successful in 23s
Details
CI / manifest (pull_request) Successful in 27s
Details
CI / json-check (pull_request) Successful in 33s
Details
4b29638947 
Refs OP#1972

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Gravity Bot commented 2026-03-16 04:15:41 +00:00
Author
Collaborator
Copy link

PR Review

Verdict: Acceptable

This review was conducted from a fresh context against the current branch state (14 files, ~1720 lines changed). First review cycle — no prior reviews or followups.

Findings

Severity Count
Blocking 0
Important 1
Minor 1
Nitpick 0

Test Results

  • Tests: N/A — generic ecosystem, no test runner detected
  • Coverage: N/A
  • Smoke Tests: N/A — no Docker services or application endpoints

Human Review Items

1 item flagged for human review.

Scope

  • 14 files changed (5 new, 7 modified, 2 deleted)
  • All documentation and configuration — no source code changes
  • Packages affected: gt-writing-quality (new), domain/frontend/taste-design (new), domain/frontend/aesthetic-direction (new), domain/methodology/output-completeness (new), gr-implementation, gt-project-setup, gt-wiki, manifest.json, gravity-skills.tsv

Key Findings

  1. [IMPORTANT] Incomplete acceptance criteria — gt-journal not updated (OP#1984)
  2. [MINOR] /humanize trigger listed but not registered — gt-writing-quality/SKILL.md

configs/claude/skills/gt-writing-quality/SKILL.md

  • [IMPORTANT] Lines 1729-1733: OP#1984 ("Update gt-wiki and gt-journal to reference gt-writing-quality") is listed as complete in the PR description, but gt-journal contains no reference to gt-writing-quality. The gt-wiki skill was correctly updated (old Writing references replaced with gt-writing-quality in 3 locations), but gt-journal has no corresponding change. The gt-writing-quality integration table lists gt-journal as an integration point (Applies quality rules to multi-line journal comments), but without a reciprocal reference in gt-journal, the journal skill won't invoke writing quality checks on its prose output.

  • [MINOR] Line 1731: The trigger conditions section lists /humanize as a valid slash command, but gravity-skills.tsv only registers /writing-quality as the trigger for this skill. After the old Writing/ skill directory is removed (correctly added to manifest.json cleanup), the /humanize command won't route to anything. Either register /humanize as an alias in gravity-skills.tsv or remove it from the trigger conditions.

HUMAN REVIEW

HUMAN REVIEW: The gt-journal omission (OP#1984) may be intentional — the old Writing skill was never referenced by gt-journal either, so there was nothing to "update." If the intent of OP#1984 was to add a new integration rather than update an existing reference, then the gt-journal side of the task is genuinely incomplete. If gt-journal was included in the task description optimistically but writing quality enforcement for journal entries is deferred, the task description should be clarified. Please verify the intended scope.

What Looks Good

  • Clean skill decomposition: aesthetic-direction (creative "what/why") vs taste-design (engineering "how") is a well-designed separation of concerns
  • Output-completeness skill is focused and actionable — bans specific placeholder patterns rather than vague guidelines
  • gt-writing-quality properly preserves Content Hygiene / Footer Normalization from the old Writing skill
  • manifest.json correctly handles both new symlinks and old directory cleanup (~/.claude/skills/Writing/)
  • Schema documentation for the taste: section is thorough with validation rules, examples, and parameter descriptions
  • Source attributions are present and specific on all new skills
  • Cross-references between skills are consistent (gt-wiki ↔ gt-writing-quality, gr-implementation → output-completeness, taste-design ↔ aesthetic-direction)
## PR Review **Verdict: Acceptable** This review was conducted from a fresh context against the current branch state (14 files, ~1720 lines changed). First review cycle — no prior reviews or followups. ### Findings | Severity | Count | | -------- | ----- | | Blocking | 0 | | Important | 1 | | Minor | 1 | | Nitpick | 0 | ### Test Results - **Tests:** N/A — generic ecosystem, no test runner detected - **Coverage:** N/A - **Smoke Tests:** N/A — no Docker services or application endpoints ### Human Review Items 1 item flagged for human review. ### Scope - 14 files changed (5 new, 7 modified, 2 deleted) - All documentation and configuration — no source code changes - Packages affected: `gt-writing-quality` (new), `domain/frontend/taste-design` (new), `domain/frontend/aesthetic-direction` (new), `domain/methodology/output-completeness` (new), `gr-implementation`, `gt-project-setup`, `gt-wiki`, `manifest.json`, `gravity-skills.tsv` ### Key Findings 1. [IMPORTANT] Incomplete acceptance criteria — gt-journal not updated (OP#1984) 2. [MINOR] `/humanize` trigger listed but not registered — `gt-writing-quality/SKILL.md` --- ### configs/claude/skills/gt-writing-quality/SKILL.md - **[IMPORTANT]** Lines 1729-1733: OP#1984 ("Update gt-wiki and gt-journal to reference gt-writing-quality") is listed as complete in the PR description, but gt-journal contains no reference to `gt-writing-quality`. The gt-wiki skill was correctly updated (old `Writing` references replaced with `gt-writing-quality` in 3 locations), but gt-journal has no corresponding change. The gt-writing-quality integration table lists gt-journal as an integration point (`Applies quality rules to multi-line journal comments`), but without a reciprocal reference in gt-journal, the journal skill won't invoke writing quality checks on its prose output. - **[MINOR]** Line 1731: The trigger conditions section lists `/humanize` as a valid slash command, but `gravity-skills.tsv` only registers `/writing-quality` as the trigger for this skill. After the old `Writing/` skill directory is removed (correctly added to manifest.json cleanup), the `/humanize` command won't route to anything. Either register `/humanize` as an alias in `gravity-skills.tsv` or remove it from the trigger conditions. ### HUMAN REVIEW **HUMAN REVIEW:** The gt-journal omission (OP#1984) may be intentional — the old Writing skill was never referenced by gt-journal either, so there was nothing to "update." If the intent of OP#1984 was to *add* a new integration rather than update an existing reference, then the gt-journal side of the task is genuinely incomplete. If gt-journal was included in the task description optimistically but writing quality enforcement for journal entries is deferred, the task description should be clarified. Please verify the intended scope. ### What Looks Good - Clean skill decomposition: aesthetic-direction (creative "what/why") vs taste-design (engineering "how") is a well-designed separation of concerns - Output-completeness skill is focused and actionable — bans specific placeholder patterns rather than vague guidelines - gt-writing-quality properly preserves Content Hygiene / Footer Normalization from the old Writing skill - manifest.json correctly handles both new symlinks and old directory cleanup (`~/.claude/skills/Writing/`) - Schema documentation for the `taste:` section is thorough with validation rules, examples, and parameter descriptions - Source attributions are present and specific on all new skills - Cross-references between skills are consistent (gt-wiki ↔ gt-writing-quality, gr-implementation → output-completeness, taste-design ↔ aesthetic-direction)
fix(skills): add gt-writing-quality reference to gt-journal
All checks were successful
PR Validation / validate-branch (pull_request) Successful in 5s
Details
CI / json-check (pull_request) Successful in 10s
Details
CI / lua-check (pull_request) Successful in 11s
Details
CI / manifest (pull_request) Successful in 9s
Details
CI / security (pull_request) Successful in 11s
Details
PR Validation / validate-release-pr (pull_request) Successful in 6s
Details
CI / sast (pull_request) Successful in 18s
Details
0110637517 
gt-journal now references gt-writing-quality for multi-line comments
(preflight/postflight reports) containing prose longer than 3 sentences.
Adds integration note in preflight template section and outbound skill
reference table in Integration Points.

Closes OP#2239

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Gravity Bot commented 2026-03-16 04:22:12 +00:00
Author
Collaborator
Copy link

PR Review Followup

Responding to review posted on 2026-03-16 (review comment).

Triage Summary

Severity Count Action
Blocking 0 —
Important 1 Fixed in this branch
Minor 1 Deferred to v1.28.0
Nitpick 0 —
Human Review 1 Flagged — requires human judgment

Fixes Applied

The following findings have been resolved on this branch:

  1. [IMPORTANT] gt-journal missing gt-writing-quality reference (OP#1984 gap) — fixed in 0110637
    • OP#2239: fix: add gt-writing-quality reference to gt-journal (OP#1984 gap)
    • Added writing quality check note to preflight multi-line template
    • Added "Skills referenced by journal" integration table with gt-writing-quality entry

Deferred to v1.28.0

The following items have been tracked for the next version:

  1. [MINOR] /humanize listed as trigger in gt-writing-quality but not registered in gravity-skills.tsv
    • OP#2241: fix: remove unregistered /humanize trigger from gt-writing-quality

Human Review Items (Unchanged)

These items require human judgment and were not auto-triaged:

  1. HUMAN REVIEW: The gt-journal integration scope may have been intentionally minimal — the old Writing skill was never referenced by gt-journal either. The fix applied here adds a lightweight reference for multi-line comments containing prose. Verify this aligns with the intended integration depth.

Test Results

  • Tests: N/A — generic ecosystem, no test runner
  • All fixes verified: Yes (commit pushed, PR updated)

This followup was generated by the gr-pr-followup skill.

## PR Review Followup Responding to review posted on 2026-03-16 ([review comment](https://git.bros.ninja/mike/kickstart.nvim/pulls/52#issuecomment-955)). ### Triage Summary | Severity | Count | Action | | -------- | ----- | ------ | | Blocking | 0 | — | | Important | 1 | Fixed in this branch | | Minor | 1 | Deferred to v1.28.0 | | Nitpick | 0 | — | | Human Review | 1 | Flagged — requires human judgment | ### Fixes Applied The following findings have been resolved on this branch: 1. **[IMPORTANT]** gt-journal missing gt-writing-quality reference (OP#1984 gap) — fixed in `0110637` - OP#2239: fix: add gt-writing-quality reference to gt-journal (OP#1984 gap) - Added writing quality check note to preflight multi-line template - Added "Skills referenced by journal" integration table with gt-writing-quality entry ### Deferred to v1.28.0 The following items have been tracked for the next version: 1. **[MINOR]** `/humanize` listed as trigger in gt-writing-quality but not registered in gravity-skills.tsv - OP#2241: fix: remove unregistered /humanize trigger from gt-writing-quality ### Human Review Items (Unchanged) These items require human judgment and were not auto-triaged: 1. **HUMAN REVIEW:** The gt-journal integration scope may have been intentionally minimal — the old Writing skill was never referenced by gt-journal either. The fix applied here adds a lightweight reference for multi-line comments containing prose. Verify this aligns with the intended integration depth. ### Test Results - **Tests:** N/A — generic ecosystem, no test runner - **All fixes verified:** Yes (commit pushed, PR updated) --- _This followup was generated by the gr-pr-followup skill._
Mike Bros approved these changes 2026-03-16 04:25:56 +00:00
Mike Bros merged commit 7727245da4 into master 2026-03-16 04:26:36 +00:00
Sign in to join this conversation.
Reviewers
No reviewers
Mike Bros
Labels
Clear labels
No items
No labels
Milestone
Clear milestone
No items
No milestone
Projects
Clear projects
No items
No project
Assignees
Clear assignees
No assignees
2 participants
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference
mike/kickstart.nvim!52
No description provided.