Skip to content

CHANGELOG.md

Latest commit

 

History

History
46 lines (32 loc) · 3.07 KB

CHANGELOG.md

File metadata and controls

46 lines (32 loc) · 3.07 KB
Raw

Changelog

Plugin-level releases. Per-skill versions live in VERSIONS.md.

1.0.1 — 2026-08-21

Documents the npx install path. No skill content changed — this worked at 1.0.0, it just wasn't written down anywhere.

  • npx skills add zstvns/design-skills --all installs all 8 skills into .agents/skills/ via the skills CLI, wiring up every detected agent (Claude Code, Codex, Cursor, Amp, Cline, Antigravity, and more) and writing a skills-lock.json. Verified end to end against the live repo: all 8 discovered with correct descriptions, every references/ and evals/ file intact after install.
  • README and INSTALL.md now lead with npx for non-Claude-Code hosts and demote the manual clone-and-symlink to the "manage it yourself" option.
  • Documented that re-syncing is safe: these skills never write state inside their own folder, so an update that replaces the skill files can't destroy a user's brand context, strategy record, or logo assets. (This is the failure mode that cost makerskills its users' archives — worth stating as a deliberate property rather than luck.)

1.0.0 — 2026-08-07

Initial public release. Eight skills, installable as a Claude Code plugin or symlinked into any Agent Skills host.

Skills

Skill Version
brand 1.26.1
brand-strategy 1.12.1
design-principles 1.11.1
creative-direction 1.14.2
logo-design 1.11.2
web-design 1.7.1
collateral-design 1.11.1
motion-design 1.6.1

Packaging

  • Added .claude-plugin/plugin.json and .claude-plugin/marketplace.json — installable via /plugin marketplace add zstvns/design-skills
  • Added INSTALL.md documenting the (entirely optional) recommended connections: Quiver API, Figma MCP, Adobe MCP, Paper MCP, and Flora or Higgsfield for generative rich media
  • Rewrote README.md with an install path, a five-minute getting-started, an intent → skill routing table, and a roadmap

Documentation

  • Documented the full context-document architecture. The library reads and writes four .agents/ documents plus a logo asset directory; only brand.md had been documented, while design.md — which web-design and collateral-design both hard-gate on — was not mentioned anywhere.

Cross-reference correctness

  • Removed every routing reference to a skill that doesn't exist yet. Shipped skills previously handed off to email-design, social-design, art-direction, brand-naming, and design-critique as though they were installed.
    • email-design / social-design → collateral-design, which already owns social posts, OG images, and email graphics as fixed formats
    • art-direction → creative-direction, which owns asset direction until a v2 art-direction skill can assume it
    • brand-naming → described as a separate naming exercise rather than a skill handoff
    • design-critique → prose reworded to "a later critique"; Related Skills entries annotated (planned — not yet shipped) so the roadmap signal survives without implying availability