jordanhubbard
- 38 skills
- 0 followers
- 7 hours ago last updated
- ▌ Responsible Vibe · jordanhubbardStructured development workflows for AI-assisted coding. Use when starting new features, fixing bugs, following TDD, refactoring code, or any development task that benefits from planning and structure. Activate it when users mention to build, enhance or fix code.
- ▌ Mac CLI · jordanhubbardHow the mac CLI is actually shaped — object groups, the admin re-parenting, and the verbs that are not what you would guess. Read this before running mac commands.
- ▌ Judgement · jordanhubbardHub-owned checklist for judging task-lifecycle gate quality and intervening on process defects. Read by the hourly judgement process and by any operator inspecting why the hub stopped a task, held an agent, or redeployed the fleet.
- ▌ Cut A Release · jordanhubbardThe documentation work that must land before a release is tagged — capture a capabilities deck pinned to the release commit, publish it to Google Slides, link it, and re-verify the docs gates. Use when every release gate is green and the next step is tagging a version and creating the GitHub release.
- ▌ Setup Mac Fleet · jordanhubbardUse when a user asks to set up, bootstrap, deploy, access, or configure mac agents or a fleet — including how to deploy onto bare-metal, virtual-machine, containerized, or provider-managed HGX agents. Runs the first-time setup wizard, writes a home-scoped multi-fleet registry, selects the right supervisor (systemd / launchd / supervisord) and SSH transport (direct / HGX / Tailscale / Headscale / bastion ProxyJump) per host type, and keeps fleet-specific data out of Git.
- ▌ Agentbus Context · jordanhubbardRead the fleet's broadcast bus BEFORE you start a task — has this work already landed, has the trunk moved under you, is a peer holding the file you are about to edit. Consulting the bus is step zero of starting work, not an optional extra.
- ▌ Record User Directed Work · jordanhubbardTurn user direction, corrections, and discovered defects into durable ledger tasks before implementing them. Use whenever you change this repository or the fleet in response to a conversation, or find follow-up work while executing.
- ▌ Mac Agent Terminal Timeout · jordanhubbardUse when a MAC fleet agent task fails with a terminal:timeout tool_error, or when running long-lived MAC-repo operations (contract tests, bootstrap, large git operations) to avoid timing out mid-task.
- ▌ Agentfabric Overview Presentation · jordanhubbard bundleRegenerate or extend the AgentFabric overview document pair (20-slide deck plus prose narrative) from this authoring package. Read before editing any file in docs/presentations/agentfabric-overview/.
- ▌ Nvidia Inference Multimodal · jordanhubbardGo beyond text: understand images (vision) and generate images via the NVIDIA inference hub through the in-mac router. Use when a task involves looking at an image/screenshot or producing an image — no local GPU needed (routes to hosted models).
- ▌ Kslug Nightly News · jordanhubbardProduce the KSLUG Nightly News broadcast transcript for Slack #localnews from collector wire copy. Use when the host runner asks for tonight's KSLUG show.
- ▌ Releasing · jordanhubbard bundleCut a NanoLang release. Defines the ordered gates, and indexes the child skills that own each one. Use before running make release, make release-minor, or make release-major, and whenever a release is being prepared by hand.
- ▌ Regenerate Nanolang Developer Overview · jordanhubbard bundleRegenerate and refresh the NanoLang developer presentation and narrative from current language, compiler, NanoISA, NanoVM, and release evidence.
- ▌ Reading Nanolang · jordanhubbardRead, understand, and write current NanoLang (.nano) source correctly. Use this whenever you inspect, review, or generate NanoLang, including src_nano/ and NanoLang modules. Covers current calls, equal-precedence operators, local inference, optional else, project shadow policy, safety boundaries, and my first-person voice.
- ▌ Mac Task Tracking · jordanhubbardTrack issues and work items for this repository using the MAC hub task ledger (`mac task`). Use this whenever you need to find available work, file a follow-up issue, claim a task, or close one. NanoLang uses `mac task` — NOT bd/beads, NOT TodoWrite/TaskCreate, NOT markdown TODO lists.
- ▌ Release Readiness · jordanhubbardInspect GitHub issues and pull requests before a NanoLang release, scope them to the release in progress, and close stale work only with an explicit, auditable annotation.
- ▌ Roadmap Execution · jordanhubbardPlan and execute repository product work through docs/ROADMAP.md. Use this before starting features, refactors, optimization programs, or other work with multiple deliverables. Add ordered checkboxes first, execute from top to bottom, and mark only verified work complete.
- ▌ Session Completion · jordanhubbardLand the plane at the end of a work session: run quality gates, update task status, and — critically — commit AND push so no work is stranded locally. Use this whenever you are wrapping up changes in this repository. Work is NOT complete until `git push` succeeds.
- ▌ Releasing Presentation · jordanhubbard bundleBring the developer deck and narrative current before a NanoLang release is tagged, and publish them afterwards. Use when make release-docs-check reports the presentation stale or disagreeing with its specification.
- ▌ Nanovm Opcode Debugging · jordanhubbardDiagnose NanoVM opcode, stack, value, and FFI failures with the optional NANO_VM_TRACE instrumentation.
- ▌ Releasing Documentation · jordanhubbard bundleBring README, CONTRIBUTING, the user guide and the roadmap current before a NanoLang release is tagged. Use when make release-docs-check names one of them, or when preparing a release by hand.
- ▌ Persona Template · jordanhubbard bundleReference template for creating new Loom agent personas. Use when adding a new role to the org chart, defining a new agent's responsibilities, or bootstrapping a persona SKILL.md from scratch. Copy this file, replace the placeholder sections, and validate with loomctl.
- ▌ Ceo · jordanhubbard bundleExecutive decision-maker who processes escalated beads, resolves agent deadlocks, allocates resources across the org chart, and produces weekly executive summaries. Use when a bead is stuck at the top of the escalation chain, agents disagree on priority or approach, resources need reallocation, or a strategic status report is needed. Handles approve/deny/reassign/cull decisions, org health monitoring, and cross-team coordination via loomctl.
- ▌ Cfo · jordanhubbard bundleFinance leader who monitors LLM provider spend, forecasts token costs, enforces budget thresholds, and recommends model selection changes to reduce waste. Use when provider costs are rising unexpectedly, a budget alert fires, a workflow needs cost optimization, or a spend forecast is required before scaling a project. Tracks cost-per-bead, flags repeated failed calls, and updates provider configuration via loomctl.
- ▌ Cto · jordanhubbard bundleMakes final-call architecture decisions, triages complex technical escalations, selects technologies, and sets engineering standards across the org. Use when a technical problem exceeds the Engineering Manager's scope, when architecture needs a binding decision, when technology choices have long-term implications, or when engineering standards need to be defined or updated.
- ▌ QA Engineer · jordanhubbard bundleValidates code correctness, discovers edge cases, and enforces quality gates before shipping. Use when you need to write test plans, run regression suites, diagnose flaky tests, analyze failure patterns, or verify that a fix addresses the root cause rather than just the symptom. Handles test plan creation, integration testing, edge case discovery, quality metrics tracking, and regression detection across the Loom codebase.
- ▌ Web Designer · jordanhubbard bundleDesigns user interfaces with clear information hierarchy, intuitive interaction patterns, and accessibility-first layouts. Use when creating wireframes, evaluating UX flows, auditing visual hierarchy, designing responsive layouts, or improving usability of existing Loom interfaces.
- ▌ Code Reviewer · jordanhubbard bundleReviews pull requests and code changes for correctness, security vulnerabilities, race conditions, performance regressions, and architectural fit. Use when a bead involves code review, security audit, merge approval, or when an agent needs a second pair of eyes before shipping. Catches bugs, enforces best practices, and provides actionable fix suggestions with concrete remediation steps.
- ▌ Decision Maker · jordanhubbard bundleBreaks deadlocks when Loom agents disagree or can't reach consensus. Analyzes competing positions, weighs trade-offs by reversibility and impact, and issues a binding decision with documented rationale. Use when agents are stuck, deadlocked, in disagreement, or a tie-breaker is needed.
- ▌ Devops Engineer · jordanhubbard bundleConfigures CI/CD pipelines, debugs build failures, automates deployments, and monitors infrastructure health in Loom projects. Use when a build fails, a pipeline needs setup, a deployment is stuck, containers need orchestration, or monitoring and alerting need configuration.
- ▌ Product Manager · jordanhubbard bundleStrategic product agent that triages customer feedback beads, writes actionable user stories with acceptance criteria, prioritizes work by user impact, and maintains roadmap alignment. Use when new customer feedback arrives, when features need prioritization, when user stories need writing, when roadmap drift is detected, or when weekly product status reports are due. Handles feature gap analysis, backlog grooming, and cross-team coordination through Loom meetings.
- ▌ Project Manager · jordanhubbard bundleTracks bead progress, manages release timelines, and resolves cross-team dependency blockers using loomctl and the Loom org chart. Use when beads are slipping, releases need coordination, or blocked dependencies threaten downstream work. Handles milestone planning, critical path analysis, risk identification, and status reporting across the engineering org.
- ▌ Housekeeping Bot · jordanhubbard bundleAutomated codebase maintenance agent that removes dead code, prunes stale branches, updates outdated dependencies, fixes lint violations, and cleans orphaned test fixtures on a recurring schedule. Use when the repository accumulates technical debt, when dependencies fall behind, when CI reports lint failures, or when stale branches and unused artifacts need cleanup. Invoked by scheduled sweeps, manual hygiene requests, or post-release cleanup passes.
- ▌ Engineering Manager · jordanhubbard bundleManages engineering ICs by triaging blocked beads, running oversight loops, reviewing architecture, and coordinating cross-agent work in Loom. Use when beads are stuck, an IC needs help, architecture decisions require review, or engineering status and velocity need assessment.
- ▌ Documentation Manager · jordanhubbard bundleCreates and updates README files, API documentation, changelogs, architecture docs, and user guides for Loom projects. Use when docs are outdated, a feature needs documenting, a changelog entry is missing, or code changes require documentation updates.
- ▌ Web Designer Engineer · jordanhubbard bundleImplements accessible, performant web interfaces from design specs and requirements using HTML, CSS, and JavaScript. Use when building UI components, optimizing page load performance, fixing accessibility issues, or converting design mockups into production-ready responsive code within Loom projects.
- ▌ Remediation Specialist · jordanhubbard bundleDiagnoses stuck agents, systemic blockers, and self-healing failures across the Loom orchestration layer. Use when an agent is looping, a recovery sweep fails to clear stalled beads, dispatch logs show repeated errors, or the organization itself is malfunctioning. Performs root cause analysis, loop detection, recovery pattern repair, and threshold tuning for Loom's agent infrastructure.
- ▌ Public Relations Manager · jordanhubbard bundleManages all external-facing communication for the Loom project including GitHub issues, release notes, community responses, and status pages. Use when drafting public announcements, triaging community feedback into beads, writing release notes from code changes, or responding to external bug reports. Translates internal technical context into clear public communication.