nthplusio-coding-agent

Plugin nthplusio-coding-agent from marketplace nthplusio in nthplusio/skills. Commands read from the repo's manifest — yesterday.

Skills (25)

• good-css

  Modern CSS techniques that replace breakpoint ladders, wrapper elements and scripts. Use whenever you write, edit or review styles in any form, including plain CSS, Tailwind classes, StyleX, CSS-in-JS and inline styles, and whenever you build or restyle a page or component, even if the user never me

• agent-handoff

  Writes actionable handoffs with a state-checking preflight, essential context, and dependency-ordered actions with completion criteria. Use only when explicitly asked to write a handoff, brief the next session, or package work for a subagent or person. Not for conversation summaries, status or stand

• auditing-business-rules

  Audits business-rule consistency, code conformance, and proposed changes against a versioned policy registry, with evidence and explicit gaps. Use when asked to audit business rules, assess a specification's policy impact, or review a code change against approved rules. Does not approve policy or cl

• business-rules-catalogue

  Derives code-enforced business rules, has a second reader check them, and builds an explorer for business owners to review and export corrections. Use only when explicitly asked to catalogue business logic, find where it lives, or refresh the rules explorer. Not for explaining one function, reviewin

• configuring-reviews

  Configures repository-specific PR review references, execution preferences and ticket-policy pointers over portable defaults. Use when setting up or tailoring code reviews for a repository, or changing what reviewers check. Not for running a review, installing a tracker or changing lifecycle automat

• dashboard

  Renders a producer-owned status file as a self-contained HTML page for non-technical readers and refreshes it at the user's chosen destination. Use for a run dashboard, thread status page, subagent progress view, or a coordinating skill's report. The producer determines status; this skill renders th

• designing-questionnaires

  Researches requirements, coordinates category research when needed, and reviews feedback through a concise questionnaire with a JSONL planning handoff. Can post owner-approved asks to a team chat channel (Microsoft Teams, Slack or another) and watch it for answers. Use when asked to collect team fee

• explain-to

  Fits explanations to a named reader and resolves audience/output settings for caller skills, reusing known choices. Use to explain to a PM or customer, teach a concept, write an overview for someone outside the code, or revise an explanation that is too technical or vague. Not for one-line facts, co

• harness-discovery

  Discovers agent capabilities and repository ticket conventions, saving sourced facts by harness. Covers sessions, subagents, wake-ups, skills, instructions, questions, models, and HTML destinations. Use when pstack-sync, dashboard, or another caller needs discovery, or for /harness-discovery. For on

• maintaining-business-rules

  Captures business decisions from conversations and documents into a versioned, owner-approved rule registry, preserving proposals, sources, and history. Use when explicitly asked to capture or maintain business rules, or when a project's enabled specification workflow calls for it. Do not monitor or

• meeting-index

  Indexes recorded meetings across a period into shared themes, a searchable ledger, decisions, owners, and risks. Reads Plaud or another recordings-and-notes source. Use only when explicitly asked to index meetings, recap recordings, or digest calls across a period or topic. Not for a single-meeting

• optimize-ci

  Evaluates, designs, and improves a repository's CI architecture for trustworthy feedback, runner cost, and maintainability while preserving merge and release contracts. Use when the user asks to audit or speed up CI, reduce CI cost, redesign a pipeline, or design CI for a repository without an exist

• picking-skills

  Opens a reusable searchable skill picker and copies an exact-source loading instruction. Use for submitted /skills messages, a <skill> placeholder anywhere in a message, or requests to browse and choose skills.

• pr-detail

  Produces an HTML retrospective on one PR with its timeline, reviewer feedback and outcomes, and an evidence-based execution score. Use only for explicit requests about how a PR went, its review process, or why it took so long. Not for teaching the code, judging an open PR's merge readiness, or recap

• pr-digest

  Builds a themed HTML recap of PRs merged to a branch over a period, optionally including ready-for-review PRs, or an open-PR inventory including drafts with no time window. Use only for explicit requests for what shipped this week, a merged-PR digest, or what's open or in the backlog. Not for one-PR

• pr-review-round

  Reviews every non-draft PR against a base branch, oldest first, with risk-scaled depth and findings verified at the current head. Reports readiness; approves, dismisses, or merges only within explicit authority. Use only when asked for a review round, queue triage, or to clear the review queue. Not

• pr-teacher

  Teaches one PR through an interactive HTML artifact with the problem, mechanism, changed-file map, operable example, and grounded self-check. Use only when asked to explain a specific PR, walk through its change, or quiz the reader on it. Not for a merged-period recap, a review verdict, or one narro

• pstack-sync

  Syncs pstack or another configured set from upstream, including skill and subagent dependencies, and replays harness adaptations. Use only when explicitly asked to sync, update, or install a managed set for this harness. Not for an unrelated single skill, a repository without a set file, or upstream

• release-brief

  Writes non-technical release briefs from verified production evidence, grouped by consequence and linked to tickets and PRs, as HTML or Markdown. Use only for explicit release-brief, board-update, or client/stakeholder shipped-work requests. Not for general recent-change recaps, changelogs, technica

• speak-clearly

  Write human-facing prose in Google developer style — named actors, condition-first sentences, timeless tense, no filler. Use when drafting docs, READMEs, comments, commit messages, PR descriptions, error messages, or UI copy; when editing existing prose; or when asked to make writing clearer, tighte

• tracking-projects

  Keeps a project's running record of work done and remaining, with lanes, owner decisions, research answers, a Done table with checks and CI, and an overview visual. Each decision gives the owner the question, context, sources, recommendation, why, and what it holds up. Use when asked to start or upd

• visualize-spec

  Turn a spec into a throwaway HTML page — problem, outcome, workflow, pieces and seams, data layer, file layout, tests, and validation plan — themed to the repository and opened locally or published wherever the agent can, so a developer can check it delivers what they want.

• work-proposal

  Drafts client proposals with code-grounded effort, agreed commercial terms, and explicit unknowns, as HTML, PDF, or email-ready Markdown. Use when asked to scope and price work, quote a phase or epic, or write a proposal or SOW. Not for estimating one ticket, internal sprint planning, or timing a si

• working-tickets

  Maintains hosted or file-based tickets under repository policy throughout execution, including comments, status, title, body, and tags. Use when assigned a ticket, resuming one, or creating or picking one up to work on. For a parent or epic, plans and works every in-scope child in dependency order.

• writing-in-voice

  Drafts and revises writing using a personal or project voice profile, learns from corrections, and formats for the destination. Use when asked to write in my voice, learn my writing style, remember a writing preference, or adapt a voice-led draft for email, Teams, Slack, text messages, documents, or

Similar plugins