Lightweight provider-neutral control plane for cost-aware rubric execution, mechanical risk gates, bounded delegation and review, portable personal setup, and evidence-bearing results.
Plugin harness from marketplace studio-moser in Studio-Moser/skills-n-stuff. Commands read from the repo's manifest — 2 days ago.
This plugin runs code on your machine — its hooks fire automatically on events like tool use. The exact commands are listed below; read them before installing.
auditAnswers questions about how the Harness is actually used and where it falls short, from local transcripts, the delegation ledger, shim firings, and the component catalog across every synced machine. Use when asked which skills or rules are used or unused, which shims still fire or could retire, what
delegateDelegates bounded execution, fixed-target or independent review, or computer-use checks to a subagent or Codex through Harness routing. Use when an agent hands such work to a worker.
fleetRuns commands on the developer's other machines over SSH and enrolls or removes machines from the Harness fleet; inventory and public keys live in the private agents repo, never private keys. Run /harness:fleet to have the agent run a command, read a log, check a build, or restart a service on anoth
model-rubricCreates or refreshes this developer's user-global model-routing rubric, which decides which model does which work, at ${XDG_CONFIG_HOME:-$HOME/.config}/studio-moser/model-rubric.yml (synced through the agents repo when present). Use when asked to set up or refresh the model rubric, when it is stale,
risk-gateDecides whether repository work needs a written plan, recovery point, or independent review under Harness risk rules. Use before work touching security, money, persisted data, a public contract, several repositories, no testing seam, or more than one context window, or when the user explicitly reque
setupSets up or checks Harness for this developer: agents repo, links, runtimes, and the model rubric. Use when Harness needs first-time setup or a health check.
syncReconciles and publishes this machine's portable agent configuration with the guarded Harness sync script. Use when asked to sync agent config.
command -v python3 >/dev/null 2>&1 && python3 "${CLAUDE_PLUGIN_ROOT}/scripts/zsh-glob-guard.py" || true
command -v python3 >/dev/null 2>&1 && python3 "${CLAUDE_PLUGIN_ROOT}/scripts/dispatch-model-guard.py" || true
Design exploration rounds: fan one brief out into independent HTML variants built by named skills and models, gate first drafts with fresh-context critics, picture them and place the round in a Figma file where each variant opens its live page, edit variants in the page with impeccable live, and carry what the owner wants into the next brief.
6 skills · 6 agents
Backend-agnostic project management for AI-native teams. Ingests research, triages and specs work items, coordinates risk-gated sprint execution, and syncs with GitHub Issues. Also includes an explicit single-task managed workflow (/pm:dev-task).
10 skills · 1 agent · hooks
Strategic intelligence system for product teams. Weekly strategy briefs, daily market research, and deep-dive analysis — a three-cadence intelligence system that keeps you focused on what matters.
4 skills
Governed image, video, and audio generation over the Kie.ai MCP. The models are the easy part — this adds the discipline around them: prices a batch and stops for confirmation before spending, downloads every result locally before URLs expire, and logs the prompt that produced each file. Can run exploration rounds on a Flora canvas, where every prompt and iteration stays visible and editable.
1 skill
Serve temporary private previews of local websites through the host's Tailscale, as real pages that outlive the agent session.
1 skill
Transcribe spoken-word audio from YouTube, Shorts, Instagram, TikTok, and Threads. YouTube captions fetched directly (fast, no model). Falls through to local mlx-whisper on Apple Silicon, with a headless-Playwright fallback for Threads.
1 skill
Talk like caveman. Cut filler. Keep technical details explicit.
22 skills · 1 command · 3 agents · hooks
Persistent memory for Claude Code: the next session starts where the last one left off.
29 skills · 1 MCP server · hooks
tasteskill 2.5.1: professional company websites at the figma.com and x.ai level, with one section per screen, real photographs, real product UI, effects kept for tech and gaming pages, and a real-browser checker and finish gate.
1 skill
Open-source AI job search that runs locally in your AI coding CLI: scan job portals, evaluate listings into a structured A-H report with a 1-5 score, tailor your CV, track applications
5 skills · hooks
Cross-session memory and token savings for coding agents.
6 skills · 1 agent · 1 MCP server · hooks
Package any website or local web build into a lightweight desktop app with Pake CLI.
1 skill
This page updates itself from your repo. Add the badge to your README, or get it Featured to pin it above the organic results with a followed link to your site.