Recently added

Technical and on-page SEO: audits, briefs, architecture, backlinks, and the Ahrefs/Semrush/GSC data pulls.

Customer and market research: discovery interviews, personas, voice-of-customer, reviews and PR.

Paid media by platform - Google, Meta, LinkedIn, Reddit - plus ad creative and app-store optimisation.

Marketing operations: analytics stack, experimentation, CRM and chat integrations, launch runbooks.

Post-signup revenue: activation, retention, churn, referrals, email/SMS sequences and monetisation.

Outbound and sales alignment: prospecting engines, enablement, battlecards, pipeline and revenue ops.

Content production: editing, calendars, repurposing, newsletters, video, podcasts, imagery and decks.

Generative-engine optimisation: AI Overviews, citations, answer-engine visibility and AI-traffic reporting.

Big Slick: the open-source marketing distribution for Claude. 247 skills spanning strategy and positioning frameworks, SEO and content, paid channels, CRO, GTM and prospecting, pipeline and exec reporting, and an AI-visibility (GEO) system — with swappable per-company context packs. 245 skills are vendored from MIT/Apache-2.0 upstreams; 2 first-party infrastructure skills are MIT.

Scarred universal rules on every session, /init that generates a verified AGENTS.md + permissions + worktree script, /lesson that turns incidents into review traps and then into guards, an adversarial pr-reviewer, a codebase scout.

Spec-driven dev workflow for Claude Code: plan → code → review → test → evidence, plus production bug hotfix, Knowledge Base scan, codebase Q&A, user stories, dependency map, cross-service system map, and business↔code docs. A native port of the author's private Auto Spec VS Code extension. Ships a read/sync-in-only git guard hook.

Daily-driver skills for Claude Code: the PR workflow, and whatever else earns its place.

Tasks channel — comments on your claimed work and knocks at your door as in-session <channel source="tasks"> events. Binds to one session via TASKS_SESSION; read-only over the tasks server /ws.

Ship every finished turn of this thread to your team's Open Hivemind server, secrets scrubbed on the laptop, and search the shared history from inside the agent.

Ship every finished turn of this session to your team's Open Hivemind server, secrets scrubbed on the laptop, and search the shared history from inside the agent.

Keep an AI agent from silently corrupting your database. A staging gate no agent can write past, read-only enforced by hook rather than by prompt, provenance-coordinate deduplication, and a catalogue of pipeline failures that succeed loudly while losing data.

AI 資料策展的紀律（繁體中文版）：代理繞不過的 staging 閘門、用 hook 而非提示強制的唯讀、以來源座標去重，以及一份「不會報錯的失敗」清單。

Hooks that tell the Infinitus Mac app the moment a session needs you or ends a turn — the phone hears about it at once.

Zero-key Claude Code cost tracking, per project and per branch. Budget caps with SIGTERM enforcement, reads your local ~/.claude session logs, and syncs to a hosted team dashboard (Goei).

Safeguard codebase integrity, eliminate destructive lazy truncation edits, and prevent regressions.

Hooks-only skills-directory plugin: registers hooks/hooks.json in sessions that open the ai-config checkout itself, where the marketplace plugin is not installed (ai-config#2004).

Run the repo's ship skill in background terminals, using the same agent runtime as the parent session, with fleet monitoring, escalations, and automatic Codex parent continuity after transient model-capacity stops. Requires the ship skill; it launches nothing else.

Drive one change end-to-end through a repo's pipeline — issue, spec, implementation, its own review passes, hand-off or merge — on GitHub or GitLab. Self-driven: every review pass runs as read-only subagents with adversarial verification, and nothing waits on an external reviewer.

Run a multi-agent discussion room: several agent sessions (Claude Code, Codex, Antigravity) argue one question in background terminals by a strict protocol — speaking in turn, objecting with references, closing objections mechanically — and the room writes a decision record when it converges, or an honest unresolved one when it does not.

Authority-aware orchestration for coding agents: bounded autonomy, explicit decision ownership, deterministic evidence, and mandatory fresh QA.

Authority-aware orchestration for coding agents: bounded autonomy, explicit decision ownership, deterministic evidence, and mandatory fresh QA.

Knowledge-layer skills for an Obsidian-vault wiki: capture routing, gatekeeping, lint, maintenance, queue, and project creation.

Gate-oriented spec-driven development: spec -> clarify -> implement -> reviewer -> worklog, one pull request per issue, with the review gate enforced by a hook rather than requested by prose.

The estate's eleven Claude Code hook scripts plus a shared gitleaks helper, packaged as a plugin.

The estate's core session-mechanics skills plus the attack-kitty verification agent, packaged as a plugin.

Routes daily target work after the repository-local readiness handoff passes.

Validate VAST, VMAP, and DAAST ad tags against IAB Tech Lab specs. Hosted MCP at https://vastlint.org/mcp. Auth none for public tools.

Language, tooling, and delivery-practice skills for agentic coding.

Reusable planning, handoff, verification, and workflow skills for agents.

The trust layer between AI and your database, so an agent's answers are defensible. Every join, metric, and named filter is confidence-scored and either FK-derived or human-approved via a review dashboard. Every answer ships a SQL receipt showing what ran and which model version it used. Works across Postgres / MySQL / Snowflake / BigQuery / Redshift / SQL Server / Oracle / Databricks / Trino / DuckDB / SQLite. Local-first: credentials, schema, results never leave your machine.

Kannaka Quantum — run Kannaka's wave-interference memory on real quantum hardware, and drive qBraid Lab compute + autonomous remote coding agents. Resonance recall as amplitude amplification, true quantum random numbers, and OpenQASM 3 circuits on qBraid's free simulator (default) or real QPUs (IonQ/Rigetti/IQM/AQT); plus qBraid Lab ops — environments, GPU/CPU compute provisioning, and launching/driving coding agents on remote instances over SSH. Ships a SKILL and an MCP server.

Video and motion skills: DaVinci, Premiere, After Effects, Remotion, FFmpeg.

Token and context efficiency skills. Includes 5 skills.

Desktop ops: Raycast, Everything, PowerToys, Docker, VS Code, Bruno, Tailscale.

Design and knowledge apps: Figma, Photoshop, Obsidian, Notion, Linear.

Everyday prompting habits and practical AI hygiene.

CAD-focused skills: SolidWorks, Fusion 360, CATIA, NX, Rhino, Blender.

3D DCC and game engines: Cinema 4D, Houdini, ZBrush, Substance, Blender, Unity, Unreal, Godot.

Major AI Skills - product-depth agentic SKILL.md library for Claude Code.

Skills for getting a change from a local checkout to GitHub: README conventions, the sample universe, and the configuration smoke test — plus Margot, the estate's non-author PR reviewer.

rig — an AI Quality Operating System for Claude Code, built on a quality-gated workbench. Give /rig:go a natural-language task: it classifies the task, runs it in an isolated git worktree, and checks the result against a per-task-type acceptance gate — the diff only lands in your working tree on an explicit accept. Machine sensors back the gate deterministically (secret leak, prompt-injection markers, gate tampering, destructive commands, OpenAPI schema drift), with recorded overrides as the escape hatch. Track parallel tasks with board and the read-only cockpit; measure whether the gate actually works with drill (seeded-bug reviewer detection rates) and stats (rubber-stamp detection). Compose recipes, personas, and knowledge as reusable bricks; a deterministic orchestrator (orchestrate.py) runs recipe state machines with parallel verification, cost-tier auto-routing learned from run history, A/B recipe experiments, and token/cost metering. Reaches beyond a live session via a stdlib-only MCP server, a GitHub Action, Slack/Teams notifications, Codex/Cursor native-layer adapters, and a read-only VS Code sidebar. Feedback loops close the gap between accept-time predictions and reality: production-outcome tracing, signed provenance, cross-session instinct learning with decay, and gap prescriptions that draft /rig:forge improvement requests. Japanese writing quality is first-class: an AI-smell catalog with cognitive-rhythm grounding and a deterministic prose-rhythm sensor. v2 adds the organisational layer: a common policy that stacks org -> team -> project and can only ever tighten downstream, role-based permissions over a fixed vocabulary, an approval flow with quorum, separation of duties and freshness binding to the approved commit, time-boxed waivers that turn --force into a named and expiring exception, a hash-chained signed audit ledger where edits and deletions are detectable, and conformance scoring that rolls several repositories up into a per-team view. All of it inert until a repository is bound with `rig-wb govern init`, so solo use is unchanged. v2.1 extends governance from the accept step to any stage of a recipe: a step may declare `actor` (the org role that owns it) and `human_gate` (halt until a qualified person signs off), and an org policy may impose a `stage:<id>` approval rule on a recipe that never asked for one. A parked run persists in `awaiting_approval` and resumes via `orchestrate approve`. v2.2 makes a run legible to somebody who does not already know the recipe: `new` prints the chosen recipe's flow map, each step transition shows position and what is next, and board gains a real denominator plus a whose-move-is-it column. It also measures what had only ever been asserted — `rig-wb wb context` counts what rig's own output costs the parent session, which is the part rig can actually observe.

Write HTML, render video. Compositions, Tailwind v4 styles, GSAP and runtime adapter animations, captions, voiceovers, audio-reactive visuals, and website capture for HyperFrames.

Generative UI: the agent builds the page the task needs, writes new widgets when the project calls for them, and answers the user's comments and edits over a messaging and collaboration bus.

Drive GitHub issues through planner → implementer → verifier subagents, locally on a Claude Code subscription. Steady-state cycle skill (issue-cycle), persistent verification baseline, policy-gated plan auto-approval, bounded CI auto-fix, crash recovery with WIP checkpointing and resume (not restart), backoff/retry on transient dispatch failures, a per-issue dispatch ledger with pre-advance checks and mechanical closing reconciliation, per-stage status reporting, greenfield kickoff (project-kickoff), worktree-parallel implementation with a supervisor loop, label lifecycle, repo onboarding (harness-setup), merges human-gated by default with opt-in policy-gated merge autonomy guarded by base-branch assertion, loud permission-denied escalation, post-merge verification, and an opt-in test-suite ratchet that files evidence-backed, test-only coverage issues under a CLAUDE.md policy.

Practical 1C:Enterprise development workflows for Codex.