Ecosystem Update — 2026-06-30

June 30, 2026 · generated by the ecosystem-update Claude Skill

TL;DR

One safe Quick Win shipped: the startup posture harness now warns when codex on PATH differs from the packaged Codex.app CLI (0.142.1 vs 0.142.4 today).
Today's strongest research signal is validation and evaluation discipline: MAS-Lab, RuVerBench, SWE-Together, TraceLab, and Always-OnAgents all map to existing AgentOps/evidence-gate work rather than a new orchestration layer.
Community skill catalogs keep expanding, but this setup already has skill-audit, skill-installer, plugin-era skills, read-only reviewers, route classification, completion gates, and omni-mem persistence; wholesale imports remain rejected.

Quick Wins

Item	Source	Type	Impact	Effort	Action
Codex CLI split-brain posture warning	Codex config reference, local `codex-runtime-doctor`	hook	2	1	Add a read-only PATH-vs-packaged CLI version check to existing `codex_config_posture.py`, which is already wired through `SessionStart`.

Auto-Implemented

Added CODEX_CLI_VERSION_SPLIT_BRAIN warning support to /Users/chadsimon/.codex/bin/codex_config_posture.py.
Backed up config.toml, hooks.json, all agent TOMLs, and codex_config_posture.py under /Users/chadsimon/.codex/backups/2026-06-30/.
Saved the secondary omni-mem summary for this run's Quick Win.
Verified with:
python3 -m py_compile /Users/chadsimon/.codex/bin/codex_config_posture.py
python3 /Users/chadsimon/.codex/bin/codex_config_posture.py --mode warn
python3 /Users/chadsimon/.codex/bin/codex_config_posture.py --mode check
Current posture output is ok: true with one warning: PATH /Users/chadsimon/.npm-global/bin/codex is codex-cli 0.142.1, packaged /Applications/Codex.app/Contents/Resources/codex is codex-cli 0.142.4.

Build Queue

SWE-Together-style multi-turn eval adapter (research) — arXiv:2606.29957 — Current task evals and Wren harness runs are mostly final-state checks; a multi-turn simulator could measure how many corrective turns a coding agent needs, not just final success.
RuVerBench-style judge calibration check (research) — arXiv:2606.29920 — Completion gates and review agents use LLM judgment; a bounded calibration harness could detect noisy rubric verification before relying on a reviewer as closure evidence.
TraceLab-style workload telemetry summary (research) — arXiv:2606.30560 — Existing telemetry is available, but a daily token/tool-call shape summary would make route budgets and MCP-heavy sessions easier to tune without changing the runtime architecture.
Always-On state lifecycle checklist for omni-mem/Wren (research) — arXiv:2606.30306 — The current setup has durable memory, task ledgers, provenance, and gates; the gap to study is recoverability/rollback obligations for state mutations, not more memory storage.

Research

MAS-Lab: A Specification-Driven Validation Framework for Reliable Multi-Agent Systems — Strong fit for AgentOps: separates intent/specification from operational control and pushes validation into explicit artifacts.
TraceLab: Characterizing Coding Agent Workloads for LLM Serving — Useful for tuning route budgets and tool-call overhead; no immediate runtime change.
SWE-Together: Evaluating Coding Agents in Interactive User Sessions — Good template for multi-turn task evals with corrective user feedback.
Can LLM-as-a-Judge Reliably Verify Rubrics in Agentic Scenarios? — Directly relevant to reviewer/completion-gate reliability; treat LLM judgment as noisy evidence that needs calibration.
Always-OnAgents: A Survey of Persistent Memory, State, and Governance in LLM Agents — Maps well to omni-mem and Wren state governance; best consumed as a checklist before changing memory lifecycle behavior.

Already Have

OpenAI developer docs MCP, live web search, hooks enabled, SessionStart readiness checks, UserPromptSubmit route classification, PreToolUse Bash safety guard, PostToolUse edit/failure context recording, Stop completion gate, omni-mem stop/precompact hooks, read-only planner/explorer/reviewer/validator agents, Python and TypeScript reviewers, workspace-write worker/chad-twin agents, AgentOps task contract, /auto canonical runtime, planning-gate, skill-audit, skill-installer, plugin-creator, OpenAI Developers plugin, browser/chrome/computer-use plugins, document/spreadsheet/presentation/PDF skills, conservative and review profiles, prompt telemetry off, destructive app actions disabled globally, official config schema header, Codex app packaged CLI check in manual runtime doctor, AGENTS.md byte-identical policy.

Rejected

Wholesale install of agent-skill catalogs from heilcheng/awesome-agent-skills, VoltAgent/awesome-agent-skills, ComposioHQ/awesome-codex-skills, or narumiruna/skills — overengineered; existing skill-audit plus selective install is the right primitive.
Increasing nested subagent depth toward Claude/Boris depth-5 defaults — rejected; current agents.max_depth = 1 and max_threads = 3 are deliberate local safety/cost controls.
Adding fork: true to local skills — rejected; this is Claude-specific/experimental and not a stable Codex skill field in the current official config reference.
Enabling native Codex memories globally — rejected; features.memories = false is intentional while omni-mem is the default memory system and prompt telemetry remains opt-in.
Adding universal auto-format hooks from community tips — rejected; no existing formatter script is universally safe across all projects, and the skill forbids new hook wiring that requires a new script.
Auto-upgrading or symlinking the PATH Codex CLI to the packaged app version — rejected as a Quick Win; the new posture warning is safe, but changing the installed CLI path/version needs explicit user direction.
Adopting dynamic workflow catalogs wholesale — rejected; this setup already has /auto, planning-gate, orchestrate-local, Wren, and explicit anti-overengineering gates.

Sources checked: https://github.com/hesreallyhim/awesome-claude-code, https://howborisusesclaudecode.com/, https://github.com/shanraisshan/codex-cli-best-practice, https://arxiv.org/search/?searchtype=all&query=LLM+agent+coding&order=-announced_date_first, https://arxiv.org/list/cs.MA/recent, https://developers.openai.com/codex/config-reference, https://github.com/heilcheng/awesome-agent-skills, https://github.com/VoltAgent/awesome-agent-skills, https://github.com/composiohq/awesome-codex-skills, https://github.com/narumiruna/skills Tier 2 fetched: yes Tier 3 fetched: partial official-docs check only; weekly GitHub release/toolkit crawl skipped because tier3_last_run is within 7 days Run at: 2026-06-30T06:33:13-04:00