菜单
Bring THE CURRENT computer back to ecosystem + machine-equality equivalence — clean repo-hygiene drift (dirty trees, stale branches/worktrees/stashes) and remediate every non-OK equality-matrix verdict. Report-first; destructive actions are guard-gated and opt-in.
默认使用会先检查来源的 Prompt；你也可以切换为直接命令，或下载本地副本。
决定是否安装前，请先阅读 SKILL.md，以及 SkillsMP 当前展示的配套文件。
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt；运行前请先检查来源。
npx skills add https://github.com/vamseeachanta/workspace-hub --skill ecosystem-equivalence-reconcile命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地？可下载 SkillsMP 当前能够提供的文件。
Write outbound email and external messages in Vamsee Achanta's voice — a subtle offer to help, never bold or rash claims. Load before drafting ANY email, LinkedIn/Collide reply, proposal note, or outreach sent under his name.
Save/publish analysis or computation results from ANY ecosystem repo to Hugging Face as a queryable, viewer-renderable dataset. Use when the user wants to "save results to hugging face", "publish dataset to HF", "hugging face data saving", "save analysis results", "hf dataset", "make results queryable", or "render via datasets-server API". Reshapes nested results into flat parquet tables, writes a dataset card with a viewer `configs:` block and provenance, applies license/public-vs-private routing, enforces a domain data-quality gate (faithful-to-source != correct), publishes to `aceengineer/<repo>-<projection>`, and verifies via the datasets-server API.
Clone, create, fork, configure, and manage GitHub repositories. Manage remotes, secrets, releases, and workflows. Works with gh CLI or falls back to git + GitHub REST API via curl.
正在显示 SKILL.md
基于 SOC 职业分类
| name | ecosystem-equivalence-reconcile |
| description | Bring THE CURRENT computer back to ecosystem + machine-equality equivalence — clean repo-hygiene drift (dirty trees, stale branches/worktrees/stashes) and remediate every non-OK equality-matrix verdict. Report-first; destructive actions are guard-gated and opt-in. |
| type | workflow |
| version | 1.0.0 |
| category | workspace-hub |
| last_updated | "2026-06-24T00:00:00.000Z" |
| source | internal |
| tags | ["equality","hygiene","reconcile","corrective-action","worktree","branches","multi-machine"] |
| related_skills | ["repo-ecosystem-hygiene","worktree-branch-sync-hygiene","repo-sync"] |
| freedom | low |
Run this on any computer to drive it back into equivalence with the rest of the ecosystem. It reconciles two surfaces at once:
docs/reports/machine-equality-matrix.html) that is not
OK / CONFORMS / EQUAL / PARITY.Driver: scripts/readiness/reconcile-ecosystem.sh (repo-root relative).
It is report-first — the default run mutates nothing.
The driver inherits the deny-by-default guard (scripts/lib/worktree_guard.py, #3143). It
never:
--apply --stash-dirty),Every action is classified before anything runs:
| Class | Meaning | --apply behaviour |
|---|---|---|
AUTO-SAFE | reversible / guard-approved | executed |
NEEDS-APPROVAL | destructive or judgement-dependent | printed only — you run it after review |
OPERATOR-ONLY | off-box (Windows licence probe, hardware, remote host) | printed only |
# 1. SEE the plan (read-only — safe anywhere, anytime)
bash scripts/readiness/reconcile-ecosystem.sh
bash scripts/readiness/reconcile-ecosystem.sh --json # machine-readable, for CI/dispatch
# 2. APPLY only the auto-safe subset (ff-pulls clean repos, etc.)
bash scripts/readiness/reconcile-ecosystem.sh --apply
# 3. Resolve dirty trees deliberately (recoverable stash), then re-pull
bash scripts/readiness/reconcile-ecosystem.sh --apply --stash-dirty
bash scripts/readiness/reconcile-ecosystem.sh --apply --equality
Work the NEEDS-APPROVAL list by hand — each line prints the exact command. For branch /
worktree disposal at scale on the primary, prefer the existing guarded cron
(scripts/cron/daily-cleanup.sh --dry-run then live) rather than ad-hoc deletion.
| Machine (role) | Hygiene cleanup | Notes |
|---|---|---|
dev-primary (ace-linux-1, control-plane) | report-only for branch/worktree | runs the intentional worktree-per-feature automation + owns the daily-cleanup cron; the driver auto-sets DESTRUCTIVE_OK=0 here so it never trampls live worktrees |
dev-secondary (ace-linux-2) | full auto-safe eligible | scratch/execution box |
ace-win-1, ace-win-2 (Windows) | scout-only; bug-fixes OK | licensed hosts; do max work on Linux, keep these thin |
The driver maps each non-OK verdict in this machine's column to a fix:
| Verdict | Cause | Corrective action |
|---|---|---|
STALE-CHECKOUT | dirty / behind / ahead / origin-ref > 12h | clean the tree (commit-push or --stash-dirty), git fetch origin main, re-run collect-equality.sh |
MISSING-EVIDENCE (machine dim) | no fresh report on this box | run collect-equality.sh here |
MISSING-EVIDENCE (harness:<provider>:*) | provider/ capability absent — or no Claude baseline on this host (Hermes-only box grades these by design) | install/auth the provider; on a Hermes-only host, expected |
BELOW-BASELINE solvers | licence probe emits present ≠ licensed | operator/Windows licence-probe follow-up (PR #2850 note) |
BELOW-BASELINE data_access | a required tier-1 sibling isn't cloned | clone the missing repo under the sibling root |
BELOW-BASELINE compute | under cores/ram/gpu floor | operator — hardware, or reconcile harness-config.yaml compute_floor |
DIVERGES / NO-MAJORITY skills | tracked symlink materialized as a text file (core.symlinks=false) | git config core.symlinks true; rm + git checkout -- the symlink paths |
DIVERGES / NO-MAJORITY harness / behavior | runtime/config drift | build-soul-runtime.sh + install-soul-runtime.sh, then re-collect |
DIVERGES / NO-MAJORITY scheduler | cron drift | setup-cron.sh --check, then re-collect |
DIVERGES / NO-MAJORITY (other) | often a stale peer report inflating divergence | re-collect on this box first; if it persists, reconcile the config |
UNREACHABLE / EXPECTED-DIFF / EXPECTED-DIVERGENCE | by design | no action |
Re-render and (if pushing) refresh the live link per [[machine-equality-matrix-live-link]]. A divergence that survives a clean re-collect is a real config drift, not a stale-report artifact — fix the config, don't just re-collect.
repo-ecosystem-hygiene — the read-only detector this complements (it reports; this remediates).worktree-branch-sync-hygiene — branch/worktree disposal workflow for the NEEDS-APPROVAL items.scripts/cron/daily-cleanup.sh — the guarded primary-host cron that owns at-scale auto-disposal.