ai-workspace-standards
Master configuration for Vibe Coding and Harness Engineering — workspace standards for Claude Code, Gemini CLI, and all AI tools
Install / Use
npx skills add 5throck/ai-workspace-standards --skill syncInstalls into whichever agent you are using.
Gemini Rules
Gemini CLI config
Quality Score
Category
AutomationSupported Platforms
Our assessment of ai-workspace-standards
ai-workspace-standards scores 72/100 on our quality scale, 2657th of 2,880 Automation skills we index.
Its Gemini Rules is 11 KB long, well organised into 8 sections and no code examples: a thorough specification that gives an agent plenty to work with.
It has no GitHub stars yet, so there is no community track record; judge it on its content.
Maintenance, license and trust
- The repository was last updated 6 days ago, so ai-workspace-standards is actively maintained.
- Our last check on 2026-10-18 found the source still online.
- No license is declared. By default that means all rights are reserved: you can read it, but reusing or redistributing it is not clearly permitted. Ask the author before building on it commercially.
- Its trust signals score 80/100, with 2 cautions from licensing, adoption, age or documentation. These come from repository metadata, not a code audit — read the skill file before letting an agent act on it.
ai-workspace-standards compared with similar skills
All 4 of these similar skills score higher than ai-workspace-standards; compare them before choosing.
| Skill | Score | Stars | Updated | Format |
|---|---|---|---|---|
| ai-workspace-standards (this skill)by 5throck | 72 | 0 | 6d ago | Gemini Rules |
| Agent-Reachby Panniantong | 100 | 103.9k | 2d ago | CLAUDE.md |
| Scraplingby D4Vinci | 100 | 89.9k | today | MCP Server |
| rufloby ruvnet | 100 | 79.5k | today | MCP Server |
| ui-ux-pro-maxby nextlevelbuilder | 100 | 137.2k | 12d ago | SKILL.md |
Frequently asked questions
- How do I install ai-workspace-standards?
- Run
npx skills add 5throck/ai-workspace-standards. The install tabs above show the steps for each supported agent. - Which AI agents does ai-workspace-standards work with?
- It is written for Gemini CLI, as a Gemini Rules file. Other agents that read the same format can often use it too.
- Is ai-workspace-standards safe to use?
- It declares no license and scores 80/100 on trust signals. Skills are instructions an agent will follow, so read the file before installing it and do not approve commands you do not understand.
- Is ai-workspace-standards still maintained?
- The repository was last updated 6 days ago, so ai-workspace-standards is actively maintained.
Skill content
View source on GitHubname: sync description: Runs the full project sync pipeline — lifecycle update, audit, L0→L1 publish, commit, push, and PR creation. version: 1.8.0 last_reviewed: 2026-10-08 status: active scope: common l2_propagate: true owner: pm prerequisites: Bun runtime metadata: type: process triggers: - sync - /sync - commit and push - create PR
Skill: sync
Context
Runs the full project sync pipeline (scripts/dev-sync.ts). This is the single mandatory pathway for all workspace commits — it handles lifecycle finalization, audit gating, L0→L1 template propagation, git branch creation, commit, push, and PR creation in one orchestrated flow.
When to Use
- After completing any set of file changes that should be committed, pushed, and PR'd
- When the user says "commit", "push", "create PR", "/sync", or "finish this branch"
- At the end of any execution plan row or multi-agent task
- As the final step in any workspace governance workflow (Phase 5/6 of the Harness Engineering Workflow)
Output Format
- A git branch
pr/<timestamp>-<slug>created (or reused) frommain - A commit with the task-staged files plus pipeline-generated files and a conventional commit message
- An open GitHub PR with the agent-written body (Why / What Changed / Test Plan / Security Checklist / Notes)
- Console output listing each pipeline step and its result
Execution Steps
-
Stage the task files (BEFORE invoking the pipeline):
git add <task files>. Only explicitly task-staged files plus files the pipeline itself generates (memory log, VERSION_MANIFEST, propagation output, ...) are committable. Dev-sync snapshots the working tree at start and at commit time. Any working-tree change that is neither task-staged nor pipeline-generated is EXCLUDED from the commit by default (scoped staging, T-20261001-015 soak exit; design:docs/designs/2026-09-12-dev-sync-scoped-staging-design.md) — stage everything the commit needs explicitly. To restore the legacy WARN-soak sweep, setSYNC_SCOPED_STAGING=0or pass--warn-staging. Never rely ongit add -Ato pick up your edits. Never leave unrelated dirt in the tree when syncing. -
Write the PR body (the agent writes it — never shell out to an LLM CLI):
-
Inspect the change:
git diff HEAD~1 --statandgit diff HEAD~1 --name-only(first 30 files). -
Write the body in English using EXACTLY this structure (keep all section headers, fill placeholders):
Why
[1-3 sentences: what problem does this solve and why now?]
What Changed
[concise bullet list of actual changes — be specific, not generic]
Test Plan
- [ ]
bun scripts/audit.tspasses - [ ] [add relevant manual or automated test steps]
Security Checklist
- [ ] No secrets, credentials, or API keys committed
- [ ] No
.envfiles staged (use.env.samplefor templates) - [ ] Dependencies unchanged or reviewed for new CVEs
Notes
[Breaking changes, deployment steps, or reviewer guidance. Write 'None' if not applicable.]
- [ ]
-
Save it to
<git-dir>/sync-pr-body.md, where<git-dir>isgit rev-parse --git-dir(e.g..git/sync-pr-body.md) — outside the working tree so it is never committed.
-
-
Run the sync script with the provided arguments, passing the body file:
bun scripts/dev-sync.ts --body-file "$(git rev-parse --git-dir)/sync-pr-body.md" "$ARGUMENTS"Branch behavior (dev-sync 1.27.0+; T-20261012-011 / T-20261013-005):
- On
main/masterthe pipeline creates a freshpr/<timestamp>-<slug>branch from the staged set — the PR title derives from the full branch-vs-main diff (origin/main...HEAD), never the message alone, so the review surface stays represented. - On an existing branch it commits there — unless that branch's PR is CLOSED (unmerged), which is refused: a dead branch under-represents the review surface (#1509's 45 files re-landed as a 1-file #1510). Re-branch from main and re-run.
- Prefer fresh branches for superseding deliveries: a reused branch can hit GitHub-server-side merge conflicts that a local 3-way merge reports clean (#1526 case). When
gh pr update-branchrefuses, re-branch the identical diff from current main, open a new PR, close the old one as superseded. - When
gh pr viewfails with 403/GraphQL (cloud sessions), dev-sync resolves the open PR via the REST API (1.28.0).
- On
-
The pipeline executes the following steps in order:
| Step | Name | Fatal? | Description |
|------|------|:------:|-------------|
| 0 | CWD Guard | FATAL | Verifies script runs from workspace root; exits if CWD mismatches import.meta.dir/.. |
| 1 | Language Gate | FATAL | Commit message / PR title must be English (CONSTITUTION.md S3); blocks non-English via language-guard.ts |
| 2 | Memory Session Entry | FATAL | Appends session summary (changes, decisions, open issues) to memory/YYYY-MM-DD.md |
| 2 | MEMORY.md Index Sync | FATAL | Updates memory/MEMORY.md index via sync-md.ts |
| 2.5 | scripts/README.md Generation | FATAL | Regenerates scripts/README.md via generate-scripts-readme.ts if the script exists |
| 3 | CHANGELOG.md Check | FATAL | Checks CHANGELOG.md [Unreleased] — blocks and exits if empty (agent must add entries first via /changelog or manual edit) |
| 3.6 | Deprecated Script Warnings | non-fatal | Scans SCRIPTS.md for deprecated scripts and prints warnings |
| 3.7 | L0/L1 Script Drift Check | non-fatal | Runs verify-scripts.ts --check-drift to detect drift between L0 and L1 script copies |
| 3.8 | Memory File Archival | non-fatal | Runs archive-memory.ts to archive old memory files |
| 3.9 | Spec Registry Check | FATAL (L0) | Runs audit.ts --spec-check --lifecycle-only — blocks on the spec-relevance Fail (code diff with no spec activity; ADR-0055 Stage 2) and any always-on audit Fail; stale/missing-spec stay WARN; escape hatch --spec-exempt=E1-E5 (AGENTS.md §5.1.1); substantive LLM-assisted work must reach this gate through the PM routing path (ADR-0078); a missing docs/specs/registry.json is a loud WARN (gate INACTIVE — Universal Design Gate, ADR-0074; activate via spec-register.ts; the registry seed ships to projects add-if-missing) |
| 3.95 | QA Pre-checks | non-fatal | Runs project tests (if package.json has test script) and warns if README_ko.md is missing |
| 3.95b | Typecheck Gate | FATAL (L0) | Runs scripts/typecheck.ts — tsc --noEmit over scripts/ against the zero-error baseline (scripts/helpers/typecheck-baseline.json; T-20260910-012). Root context only; L1+ contexts skip cleanly via the missing-baseline guard. Never raise the error count — fix the error instead |
| 3.96c | Session-Evidence Skill Review | non-fatal | Runs skill-session-review.ts — accumulates Observed Symptom + Evidence from the day's ## Skills Used section into memory/skill-review/ (diagnosis/candidate left empty for human triage); plus a non-fatal full skill-dependency-analysis.ts --report pass (SkillHone-inspired loop; design doc docs/designs/2026-09-06-skill-session-review-design.md) |
| 3.97 | Governance Reflection Gate | FATAL (L0) | Runs verify-adr-governance.ts --strict — blocks sync when post-cutoff Accepted ADRs lack governance-doc references (ADR-0059 Stages 2+2b: unlinked-ADR and marker-drift findings block); skipped in scaffolded projects (L0-only validator) |
| 4.5 | L0 to L1 Publish | FATAL (L0) / non-fatal (L1) | Propagates scripts, skills, commands, docs via propagate-to-templates.ts --apply; fatal only in L0 context (CONSTITUTION.md present) |
| 4.52 | Template Dependency Sync | FATAL (L0) | Runs sync-template-deps.ts --apply — aligns shared dependency versions from root package.json into templates/common/package.json and regenerates bun.lock (shared keys only; root-only deps never added, template-only deps never removed); workspace-root gated |
| 4.62 | Cascade Re-publish | FATAL (L0) / non-fatal (L1) | Re-runs propagate-to-templates.ts --apply after skill sync — heals template platform skill copies (templates/common/.claude/.gemini/.agents/skills) changed by step 4.6 within the same sync; same gating and fatality as step 4.5 |
| 4.6 | Skill Sync to Platforms | non-fatal | Runs sync-skills.ts to distribute skills to .claude/skills/, .gemini/skills/, .agents/skills/; warnings only |
| 4.7 | VERSION_MANIFEST.md Generation | FATAL | Generates VERSION_MANIFEST.md via generate-version-manifest.ts |
| 4.85 | Version-Bump Test Gate | FATAL | When the task-staged change set bumps a version (a SKILL.md frontmatter version: line or a script @version header line changed), runs bun run test:unit BEFORE the audit gate; a failing suite blocks the sync. Skipped cleanly when package.json has no test:unit script (L1+ projects) and when nothing staged bumps a version. No escape env var — fix the tests or unstage the bump (detection: scripts/helpers/version-bump.ts; user directive 2026-10-08) |
| 4.9 | AUDIT GATE | FATAL | Runs audit.ts — must exit 0 before proceeding. Includes the auto-activating gates: skill-graph drift (ADR-0060) and upgrade coverage (check-upgrade-coverage.ts --strict, ADR-0073) |
| 5 | Branch Creation | FATAL | Creates pr/<timestamp>-<slug> branch if on main/master; reuses existing branch otherwise |
| 6 | Sensitive File Guard + Scoped Staging + Commit/Push | FATAL | Guards against .pem, .key, .env, credentials.json, etc.; scoped-staging check (S0/S1 tree snapshots — WARN-lists files that are neither task-staged nor pipeline-generated; excluded by default; SYNC_SCOPED_STAGING=0/--warn-staging restore the legacy sweep), then git add, git commit, git push |
| 7 | PR Creation | FATAL | If --body-file was passed, validates it (English) and opens the PR via gh pr create --body-file; otherwise falls back to gen-pr-body.ts template, .github/pull_request_template.md, then gh pr create --fill; idempotent — updates existing PR if one already exists for the branch |
- If audit fails, fix the reported issue before re-running.
Session Evidence Duty (## Skills Used)
While the session is still open — BEFORE invoking /sync — record the skills
actually loaded this session into memory/YYYY-MM-DD.md under ## Skills Used
(one block per skill; schema in the skill-lifecycle standards, §6.6
Session-Evidence Skill Review Loop, and docs/designs/2026-09-06-skill-session-review-design.md). If the section was not filled before
sync, the dev-sync step 2 skeleton is appended as a comment for later editing.
Step 3.96c consumes this evidence automatically; it never modifies skills.
Related Skills
scripts/dev-sync.ts— Core sync pipeline implementationskills/gateguard/SKILL.md— Pre-edit quality gate (run before edits that lead to sync)CONSTITUTION.md §3— PR workflow and branch rules
PR Language Rule
All PR titles and bodies generated by this command must be written in English, regardless of the active session language. This applies to the agent-written body in step 1 and to any gh pr create / gh pr edit calls — dev-sync.ts blocks non-English bodies at the same language-guard gate used for commit messages.
Related Skills
Agent-Reach
103.9kGive your AI agent eyes to see the entire internet. Read & search Twitter, Reddit, YouTube, GitHub, Bilibili, XiaoHongShu — one CLI, zero API fees.
Scrapling
89.9k🕷️ The BEST Web Scraping framework that handles everything from a single request to a full-scale crawl! Don't be shy, join here: https://discord.gg/EMgGbDceNQ and follow here for daily tips and tricks: https://x.com/Scrapling_dev
ruflo
79.5k🌊 The original agent harness. Deploy intelligent multi-player swarms, coordinate autonomous workflows, and build conversational AI systems. Features adaptive memory, self-learning intelligence, federation, vector RAG integration, and native Claude Code / Codex / Hermes and many more Integrated
ui-ux-pro-max
137.2kUI/UX design intelligence for web, mobile, and desktop. This skill should be used when designing, building, reviewing, or fixing interfaces, including pages, components, design systems, accessibility, interaction, responsive layout, typography, color, charts, and stack-specific UI implementation.
Trust signals
From repository metadata: license, adoption, age and documentation. Not a code audit — see the Safety scan above for what the skill file itself contains.
