Skip to content

v2.7.0 — aged-out audit: remove & reframe natively superseded plugins

Choose a tag to compare

View all tags
@charlesjones-dev charlesjones-dev released this 09 Jul 16:37
· 4 commits to main since this release
v2.7.0
3a70cd4

Marketplace-wide "aged-out" audit against native Claude Code features (July 2026): native Dynamic Workflows, auto memory, /code-review / /security-review, /statusline, the Learning output style, and Tool Search. Every plugin was inventoried; superseded functionality was removed or repositioned, and staleness bugs found along the way were fixed. The marketplace now lists 14 plugins.

💥 Breaking / Removed

  • ai-learn removed entirely (was v1.1.0) — superseded by Claude Code's built-in Learning output style (/config → "Output style" → "Learning"), which provides the collaborative mentor / learn-by-doing mode natively. Lives on in git history.
  • ai-workflow → v2.0.0: /workflow-plan-phases and /workflow-implement-phases removed — superseded by native Dynamic Workflows (JavaScript runtime orchestrating subagent fan-out with per-agent token budgets). The plugin now focuses on /workflow-preflight, /workflow-ship, /workflow-principles, and /workflow-rules.

The root README has a new Deprecated & Removed section tracking retired items at a glance.

🔄 Reframed against native features

  • ai-knowledge v1.5.1 — positioned as a curated, git-versioned, team-shared knowledge base that complements native auto memory (MEMORY.md), with a new "How This Differs from Claude Code's Auto Memory" section.
  • ai-security v1.5.1 — the dated "Why Not Just Use /security-review?" section replaced with an honest complement comparison: archived OWASP-mapped reports, deployed-site dependency scanning, settings hardening, supply-chain hardening.
  • ai-statusline v1.2.4 — positioned as an enhancement over native /statusline: progress bar with color thresholds, 5h/7d rate-limit widgets, cost/duration segments, edit wizard.
  • ai-git v1.3.1 — /git-pr-codex-loop now notes that native /code-review --fix/ultra covers the same intent without external dependencies; the skill remains the right tool for Codex-standardized repos.

🐛 Fixed

  • Removed hardcoded Co-Authored-By: Claude Opus 4.6 commit trailers from /git-commit-push-pr and /workflow-ship.
  • ai-modernize v1.1.0 — created the missing modernize-auditor agent that both skills delegated to (previously a broken reference); refreshed frontier-model references.
  • ai-statusline — fixed a broken statusline-setup skill reference; refreshed stale "Opus 4.6" example labels; removed vestigial "requires Claude Code 2.1.80+" caveats.
  • ai-security — stale report-template scanner version; minimum-release-age frontmatter casing; Tool Search deferral note for Context7 MCP tools.
  • ai-geo — README version drift corrected (v1.1.0). ai-ado v1.3.2 — README now documents the shipped /ado-work-items skill.
  • Auditor agents (accessibility v1.4.1, performance v1.2.1, swift v1.0.1, security) reframed from "use when context is saturated" to unattended/scheduled domain reviews; performance-auditor typo fixes.

📚 Docs

  • Root README: alphabetized plugin table, new "Deprecated & Removed" section, corrected usage note, v2.7.0 badge.
  • CLAUDE.md: removed the nonexistent /plugins-scaffold workflow, documented plugin agents and their conventions, added a deprecation/removal policy.

Full Changelog: v2.6.0...v2.7.0

Assets 2
Loading