Ecosystem Update — 2026-04-21 (dry-run)

April 21, 2026 · generated by the ecosystem-update Claude Skill

TL;DR

Current setup is saturated against the "easy wins" surface — 167 previously-seen items and a mature 10-agent / 39-skill / 10-hook-event config mean most new signal is already covered or rejected under the anti-overengineering gate.
Zero true Quick Wins today. One candidate ("agent" default-agent field in settings.json) is plausible but unverified against official docs — parked in Build Queue until confirmed rather than auto-applied.
Highest-signal NEW items are two new hook types (agent and http) from the 2026 early-features cheatsheet and a meta-review arxiv paper (2604.18071) surveying architectural design decisions across 70 agent harnesses — worth a focused read against our own runtime.

Quick Wins

None today.

Rationale: every candidate that scored Impact≥2 / Effort=1 either (a) requires a new script file (prohibited by Quick-Win hard limits), (b) introduces a frontmatter field not yet verified in official Claude Code docs, or (c) is already covered by an existing primitive. A saturated surface is the expected steady-state; reporting "none" instead of force-fitting an item is the correct outcome.

Build Queue

Skill-scoped safety blocks — /careful and /freeze (skill-pattern) — shanraisshan — On-demand hook blocks embedded inside a skill definition that block destructive commands (/careful) or edits outside a target directory (/freeze). Useful for high-risk skills like drive, orchestrate-local, and audit. Build: add the two block types, wire into 2–3 high-risk skills first.
idle-timing hook (hook) — rohitg00 toolkit — Injects hidden timing context so Claude knows how long a conversation has been paused. Relevant to our autonomous /loop and /drive runs where the model currently has no signal that a 45-minute gap occurred. Build: PostCompact-adjacent hook that writes elapsed seconds into session context.
rulecheck composite (skill + agent + hook) (composite) — coleam00 cheatsheet — Autonomous rule-violation scanner that edits code, validates the fix, and notifies Slack. Partial overlap with insecure-defaults/sharp-edges plugins but the autonomous remediation leg is new.
New hook types: agent and http (hook) — coleam00 cheatsheet — Current settings.json only uses type: command (and type: prompt is in the seen list from prior runs). The agent type fires a subagent on a lifecycle event; http posts to a URL. Verify these are real in current Claude Code spec before wiring.
Observable Loop Prevention Guards (5-layer) (pattern) — affaan-m/everything-claude-code — Re-entrancy detection with throttling, tail sampling, and lazy-start logic to prevent runaway autonomous loops. Direct applicability to /loop, /drive, /orchestrate-local, /evolve. Build: python helper in ~/.claude/bin/ that inspects recent event log for re-entry signatures and aborts.
Strategic Compaction Hooks (proactive trigger) (hook) — affaan-m — We already have PreCompact, but it only fires at the context-window threshold. Proactive trigger = suggest /compact at logical task breakpoints (slice acceptance, track closure). Lower-priority since it overlaps with existing PreCompact + memory gates.
deep-dive skill (skill) — rohitg00 — DAG-based research breaking questions into parallel dependency graphs with sourced reports. Overlap with our rlm-scan + explorer subagent pattern but adds explicit dependency DAG management.
memory: project agent frontmatter (agent-pattern) — coleam00 cheatsheet — Persistent project context across runs and worktrees via one frontmatter field. Would be a Quick Win if verified; parked in Build Queue because the cheatsheet is the only source so far and I don't want to add an unsupported frontmatter field to chad-twin/chad-agent blindly.
"agent" default-agent setting in settings.json (config) — Boris's tips — Set "agent": "chad-twin" to override the default. Same verification concern as above — Boris is the creator of Claude Code so this is likely real, but worth confirming against ~/.claude/bin/claude_run expectations before applying.
VibeGuard-style rule pack (top 10, not 88) (plugin) — rohitg00 — 88 hallucination-prevention rules + 13 hooks is too heavy to adopt wholesale; cherry-pick the 10 most aligned with our Python/TypeScript workflows.
additionalDirectories in settings.json (config) — Boris's tips — Relevant only if multi-repo work becomes a pattern. Currently single-repo per session, so deferred.

Research

arXiv:2604.18071 — Architectural Design Decisions in AI Agent Harnesses — Empirical study of 70 agent projects identifying five recurring design dimensions. Worth an explicit pass against our own ~/.claude/ setup to see if we're missing a dimension or over-investing in one.
arXiv:2604.17699 — SelfHeal: Multi-Agent ReAct for Bug Repair — Fix agent + critic agent with internal rules and web search. Partial match for our implementation-checker → reviewer → validator chain; compare against their ablation to see which link we're missing.
arXiv:2604.18349 — HiGMem: Hierarchical LLM-Guided Memory — Two-level event-turn memory allowing inspection of summaries before deep retrieval. Our native ~/.claude/memory/MEMORY.md + omni-mem is two-tier but not hierarchical in the HiGMem sense. Evaluate whether an explicit "summary-then-drill-down" access pattern would reduce omni-mem roundtrips.
arXiv:2604.18240 — AJ-Bench: Agent-as-a-Judge for Environment-Aware Evaluation — Judges that actively interact with environments to acquire verifiable evidence outperform passive LLM judging. Direct applicability: our acceptance_check shell gate is a weak form of this; the paper likely has sharper patterns for reviewer/validator agents.

Already Have

claude-pace, clawd-on-desk, ccxray, encyclopedia-agentic-coding-patterns, /techdebt, /batch, /schedule, /loop, /simplify, /go, /caveman, agent-teams-env, teammate-mode-in-process, rc-remote-control, remote-flag, teleport-flag, save-task-list, triage-command, important-if conditional tag, claude-rules-subdirectory-modularization, reporecall-ast, instinct-scoring-agentshield, channels-webhook, type-prompt-hook-type, auto-dream-memory, voicemode-mcp, claude-code-hooks-mastery, claude-code-hooks-battle-tested, second-opinion plugin (covers Bouncer), trailofbits insecure-defaults + sharp-edges (partial VibeGuard overlap), skill-installer, explorer read-only pattern, context-fork agent pattern, isolation-worktree on worker/reviewer/planner, CLAUDE_CODE_AUTO_COMPACT_WINDOW=400000 env, xhigh effort level default, skills-janitor scan, skills-janitor skill

Rejected

AuraKit (46-mode fullstack skill, 23 sub-agents, 55% token-savings claim) — fails anti-overengineering gate; no one-sentence proof that our existing 39 skills + 10 agents cannot satisfy the same use cases.
moyu anti-overengineering skill — already covered by /simplify, /what-would-chad-do, and the anti-overengineering gate baked into ~/.claude/CLAUDE.md.
Axiom contract-enforced pipeline — already covered by planning-gate skill + governance runtime (/govern, auto_runtime.py, reviewer barriers).
Bouncer Gemini quality-gate plugin — functional duplicate of second-opinion@trailofbits which is already enabled.
/code-review cloud-based PR analysis — already covered by codex-branch, codex-security, review skill, and second-opinion plugin.
Technology Scout agent — "build-vs-buy evaluator" doesn't match a recurring workflow for CW work.
AutoResearch agent — ML experiment automation via tree search; wrong domain.
Global AGENTS.md cross-harness parity file — would only matter if Codex and Claude Code session pairs became a regular pattern; currently each owns its own home per policy. No pain point justifies the duplication.
Multi-Language Rules Architecture — reorganizing skills into per-language subdirectories is a structural change without a concrete recurring problem; current flat structure works.
Spinner verbs, Output Styles tuning — cosmetic; no behavior impact.
Dispatch (mobile → desktop session handoff) — platform-level feature, no local config needed.
naming skill — not a recurring workflow.

Sources checked: - https://github.com/hesreallyhim/awesome-claude-code - https://howborisusesclaudecode.com/ - https://github.com/shanraisshan/claude-code-best-practice - https://arxiv.org/search/?searchtype=all&query=LLM+agent+coding&order=-announced_date_first - https://github.com/rohitg00/awesome-claude-code-toolkit - https://code.claude.com/docs/ - https://github.com/coleam00/claude-code-new-features-early-2026/blob/main/CHEATSHEET.md - https://github.com/affaan-m/everything-claude-code

Tier 2 fetched: yes (last run 2026-04-16, >24h) Tier 3 fetched: yes (last run 2026-04-13, >7d) State file: NOT updated (dry-run) Auto-implement: NOT executed (dry-run) Run at: 2026-04-21T17:05:00Z