polaralias/homeassistant-cozylife
Instructions file CodexOpenCode
Instructions for polaralias/homeassistant-cozylife, covering repository operating guide, core rules, documentation rules, current product frame and preferred documentation sources.
polaralias/homeassistant-cozylife
Instructions file CodexOpenCode
Instructions for polaralias/homeassistant-cozylife, covering repository operating guide, core rules, documentation rules, current product frame and preferred documentation sources.
Instructions file CodexOpenCode
AGENTS.md instructions for polaralias/skills, covering agents, working rules, versioning rule, repo release rule and repo maintenance rule.
Instructions file
Claude Code instructions for polaralias/skills, a project described as: Curated library of reusable agent skills for engineering, delivery, content, design, documentation, and automation.
Skill Claude CodeCodex
Use when the user asks to convert Markdown, HTML, DOCX, EPUB, or another Pandoc-supported format; says to run a file through Pandoc; or needs predictable wrapping with explicitly approved filters. Converts files through the repository wrapper with stable defaults. Do not use for editing Word layout (DXA) or PDF-only…
Skill Claude CodeCodex
Use when the user asks for an Excel Gantt, project timeline workbook, task bars and dates, parent-child task hierarchy, or Gantt formatting changes. Produces or updates a stakeholder-ready spreadsheet from task data. Do not use for a project plan without a workbook deliverable or for general spreadsheet analysis.…
Skill Claude CodeCodex
Use when the user asks to write an agenda, plan a meeting, turn a meeting title into discussion topics, or structure objectives, timings, owners, and decisions for a session. Produces a lean working agenda or formal stakeholder agenda. Do not use for processing notes after the meeting (MPP). Shorthand AGN.
Skill Claude CodeCodex
Use when the user asks for a short LinkedIn post, company update, capability highlight, event post, launch snippet, or concise founder-style social copy. Produces brief external-facing content sized for a social post. Do not use for articles, long LinkedIn pieces, or blog-length copy (LFP). Shorthand LSP.
Skill Claude CodeCodex
Use when the user asks for a long LinkedIn post, launch article, thought-leadership piece, blog post, or substantial external story grounded in product, customer, or campaign evidence. Produces structured long-form copy with an evidence trail. Do not use for a short social update or launch snippet (LSP). Shorthand LFP.
Skill Claude CodeCodex
Use when the user asks to process meeting notes or a transcript, summarise what happened, extract decisions and actions, draft a follow-up email, or route product feedback from a meeting. Produces an internal meeting pack and only the justified follow-up outputs. Do not use to plan the meeting before it happens (AGN).…
Skill Claude CodeCodex
Use when the user asks to write release notes, announce shipped changes, turn implementation details into customer-facing updates, or combine several completed changes into a release summary. Produces concise benefits, useful technical context, and explicit source gaps. Do not use for a pre-release implementation plan…
Skill Claude CodeCodex
Use when the user asks to find meeting times, check calendar availability, spot similar existing meetings, propose up to three slots, draft a scheduling email, or handle a batch of meeting requests. Produces calendar-aware proposals and ready-to-send wording using live calendar data when available. Do not use to write…
Skill Claude CodeCodex
Use when the user asks whether an AI idea is worth pursuing, wants to shape an early concept, needs discovery questions, or wants several AI opportunities prioritised before detailed scoping. Produces a structured initiative brief and prioritisation view. Do not use once a prioritised initiative is ready for…
Skill Claude CodeCodex
Use when the user has a prioritised AI initiative and asks to test feasibility, define POC or MVP boundaries, assess production readiness, scope delivery, or prepare a proposal or decision pack. Produces a challenged, implementation-shaped scope. Do not use for early idea discovery or portfolio prioritisation (AIB).…
Skill Claude CodeCodex
Use when the user asks to turn a project brief into ClickUp, design a ClickUp hierarchy, choose tags, or configure Gantt, board, and saved views for delivery. Produces a practical ClickUp planning structure rather than a flat task dump. Do not use for a tool-neutral implementation plan (IPW) or repository-local…
Skill Claude CodeCodex
Use when the user shares messy customer feedback and asks for a problem statement, feature request, backlog candidate, product note, or RICE prioritisation. Produces a confirmed, tracker-ready item without assuming a tracker. Do not use for publishing an already-approved work package (TPU) or broad project…
Skill Claude CodeCodex
Use when the user asks for a customer implementation plan, onboarding plan, rollout plan, or setup-to-go-live delivery sequence from kickoff material and confirmed assumptions. Produces a customer-facing plan without generic project-management overhead. Do not use for engineering feature decomposition (DDD), ClickUp…
Skill Claude CodeCodex
Use when the user asks to summarise a kickoff, recap discovery or onboarding, create an implementation overview, or synthesise mixed notes, transcripts, and source files after an initial project session. Produces an evidence-backed audience-appropriate summary with traceable references. Do not use for ongoing project…
Skill Claude CodeCodex
Use when the user asks to create or refresh PROJECT.md, consolidate scattered project notes, or establish one durable source for a real project or programme. Produces canonical project context that people and agents can reuse. Do not use to package an already-current PROJECT.md for another audience (PKG) or to write a…
Skill Claude CodeCodex
Use when the user has a current PROJECT.md and asks for an audience-specific briefing, reusable project pack, handoff, system prompt, or derivative context artefact without rediscovery. Produces tailored outputs from existing canonical project truth. Do not use when PROJECT.md is missing or stale (PCB) or when current…
Skill Claude CodeCodex
Use when the user asks for a project status report, health view, launch-readiness update, risk summary, milestone review, or executive delivery report. Builds the report from current delivery signals and checks it against structured execution data and project context. Do not use to create the canonical PROJECT.md…
Skill Claude CodeCodex
Use when the user asks where to start with project work, whether an effort is a real project, where PROJECT.md lives, whether project lookup metadata is valid, or which project skill should handle the request. Inspects and routes without producing a specialist deliverable. Do not use when the required output already…
Skill Claude CodeCodex
Use when the user asks for a customer training plan, facilitator plan, learning-session design, or training built around agreed scope, objectives, UAT, or sandbox workflows. Produces a concise operational plan and a detailed facilitator version. Do not use for a generic implementation plan (IPW) or training-document…
Skill Claude CodeCodex
Use when the user asks for a Mermaid flowchart, process map, architecture diagram, system view, cleanup of existing Mermaid code, or a diagram reconstructed from notes or a screenshot. Produces code-first Mermaid and optional rendered artefacts. Do not use for a complete persistent product design system (SDS) or a…
Skill Claude CodeCodex
Use when the user asks to derive a reusable design system from screenshots, saved pages, logos, tokens, UI kits, icons, or an existing product and persist it for coding agents. Produces DESIGN.md plus a reusable design skill from real visual evidence. Do not use for a one-off diagram (MFD) or direct frontend…