agent.md

Scans Jira for stale issues, duplicates, and missing fields, then fixes what the user approves. Use when the user says the board is a mess, asks for a backlog grooming pass, wants to clean up Jira, or asks about stale/duplicate tickets before sprint planning.

The front door to this plugin. Listens to what the user needs right now — vague or specific — and routes them to the right skill. Trigger whenever the user asks "what can you do," "help me with the team/board/sprint," "what should I focus on," or any open-ended project-management request that doesn't clearly match a single skill.

Update, create, or reprioritize the team's roadmap. Use when adding a new initiative and deciding what moves to make room, shifting priorities after new information comes in, moving timelines due to a dependency slip, or building a Now/Next/Later view from Jira epics.

Closes out a 2-week sprint — reconciles planned vs. completed scope, flags spillover and scope creep, writes a retro-ready summary, and gives a heads-up on the next sprint's readiness. Use when the user says the sprint is ending, asks for a sprint report, wants retro prep, or asks what's carrying over to next sprint.

Plan the next 2-week sprint — scope work from the Jira backlog, estimate team capacity (accounting for PTO), set a sprint goal, and draft a sprint plan. Use when kicking off a new sprint, sizing the backlog against team availability, deciding what's P0 vs. stretch, or handling carryover flagged by `sprint-close`.

Generates a one-page weekly status brief from Jira — sprint burndown, what shipped, what's blocked, and the week ahead. Built for a Scrum team running 2-week sprints. Use when the user asks for a status update, standup summary, sprint check-in, or "what's the team working on."

Scores open Jira tickets by urgency, staleness, and priority to produce a ranked "work these first" list with context, drafts status-update comments, and flags unassigned or overdue issues. Use when the user asks what to work on next, wants a triage pass on the backlog, or asks "what's most urgent" or "what's overdue."

Write a feature spec or PRD from a problem statement or feature idea. Use when turning a vague idea or ticket request into a structured document, scoping a feature with goals and non-goals, defining success metrics and acceptance criteria, or breaking a big ask into a phased spec ready to become Jira epics and stories.

Diagnoses mismatches in skills, rules, or workflows and applies the smallest necessary change. Use when the user says "adapt based on this", "what should change after this?", "this keeps happening", "this failed, what should change?", "the workflow no longer fits", "the constraints changed", or asks what skill, rule, doc, eval, memory, or process should change.

Explores user intent, requirements, and design before implementation. Use when the user asks "what should we do about X?", "ideas for…", "how should we approach…", or any open-ended design question before the team picks a direction.

Catches duplicated knowledge, logic, or structure that should have a single authoritative source. Use when the user asks to "find duplication", "deduplicate", "DRY this up", or requests an explicit quality review of code, docs, or data models.

Sequences work into ordered phases with dependencies and success conditions. Use for planning requests like "break this down", "roadmap", "approach", "milestones", "how should we proceed", migration planning, rollout planning, and scoped next steps.

Creates, updates, queries, and lints a structured wiki knowledge base. Use when the user says "add to wiki", "update wiki", "search wiki", "ingest this", "remember this", "save context", or asks to preserve a fact.

Writes and revises product requirements documents with structured sections covering goals, scope, requirements, and success criteria.

DEPRECATED — use to-ticket instead. Writes and revises user stories with acceptance criteria and story points in sprint-ready format.

Writes and revises Jira, GitHub, and Linear tickets — bugs, features, tasks, chores, documentation, and spikes — using type-specific templates.

Write product specifications and requirements documents. Use for tech specs, design docs, TDDs, functional or non-functional requirements, data contracts, UI specs, release specs, handoff docs, and system behavior.

Converts a video URL, local video/audio file, or existing transcript/text into structured, concept-first lecture notes.

Reads .pdf, .docx, .pptx, .xlsx, .msg, and .epub files by converting them to Markdown with markitdown, then ingests the result. Use when a file with one of those extensions is uploaded or a local path to one is given and the goal is to read, summarize, or understand its content. Also use when the user explicitly asks to "convert to markdown" or "save as markdown" for a supported file. Proactively convert on upload — do not wait for the user to ask. Do NOT use for creating or editing documents, for web URLs, for HTML files, or for plain text.

Draft, revise, and prune skills authoring rules and the Trigger, Structure, Steering, and Pruning checklist.

Surfaces structured findings across consistent quadrants, not freeform prose assessments. Use when the user says "review", "critique", "evaluate", "what's wrong with", "compare", or passes a skill, rule, doc, spec, code, diff, or pull request for assessment.

Reviews code for correctness, security, performance, and style with structured findings. Use when the user asks to review code, including code changes, diffs, pull requests, branches, or patches.

Generates production-ready backend code with auto-detected stack (Go, Node.js, Python, Java, Kotlin, Rust, and more), confirmed API contracts and DB schema before writing, and a P0–P3 quality checklist covering security, API correctness, observability, and performance. Use when the user asks to build an API endpoint, implement a service, add middleware, create a worker or queue consumer, or write a migration.

Generates production-ready SQL, migrations, and ORM code with auto-detected engine, confirmed schema and query contracts before writing, and a P0–P3 quality checklist covering SQL injection, destructive change safety, performance, and migration correctness. Use when the user asks to design a table, write a migration, add an index, write a query, or use an ORM.

Generates production-ready frontend code with auto-detected stack, interface contracts, decomposition heuristics, and a P0–P3 quality checklist covering a11y, TypeScript, SEO, performance, and error handling. Use when the user asks to build a component, create a page, implement UI, or scaffold a form.

Generates production-ready test code with auto-detected stack and framework, confirmed test plan before writing, and a P0–P3 quality checklist covering determinism, coverage, assertions, and CI readiness. One test type per invocation. Use when the user asks to write automated tests: E2E, API, integration, or load/performance.

Produces a contract-first interface spec before implementation, covering OpenAPI/Swagger, AsyncAPI, and GraphQL schemas. Use when asked to design an API contract, endpoints, schemas, request/response shapes, or "what should the API look like?"

Transforms outdoor spaces into functional landscapes using expert plant knowledge and architectural styling. Use when the user mentions "landscape design", "yard design", "garden planning", "plant selection", "backyard makeover", or "landscaping ideas".

Acts as a plain-language legal advisor for people without legal education. Use when the user asks for help with legal documents, contracts, legal questions, compliance review, or drafting legal text.

Catches speculative additions before they are built. Use when the user asks to "flag YAGNI", "is this needed?", "trim scope", or requests an explicit scope audit of code, plans, or specs.

You MUST use before installing, updating, or trusting ANY skill from ANY source. Audit SKILL.md, permissions, dependencies, prompt-injection patterns, network behavior, exfiltration risk, bundled resources and suspicious patterns.

Audit and rewrite content to remove AI writing patterns ("AI-isms"). Use this skill when asked to "remove AI-isms," "clean up AI writing," "edit writing for AI patterns," "audit writing for AI tells," or "make this sound less like AI." Supports a detect-only mode, an edit-in-place mode for files, an optional voice profile (casual / professional / technical / warm / blunt), and an iterate-to-convergence pass.

You MUST use this for branch name requests and branch actions. Create, switch, or rename Git branches using repository-aware branch naming conventions.

You MUST use this for commit-message requests and committing staged changes. Generate, improve, or apply Conventional Commit messages using staged changes and repository history.

DEPRECATED — use the wiki skill instead. Persists durable facts, decisions, and conventions to structured memory files.

Manage active work across people, agents, tasks, dependencies, blockers, and handoffs. Use for coordination requests like "manage this work", "lead this", "assign", "delegate", "track blockers", "status", "handoff", and multi-workstream execution.

Write or revise reference documentation for existing APIs. Use for REST endpoints, GraphQL operations, gRPC methods, WebSocket messages, authentication, parameters, schemas, responses, errors, limits, and executable request examples.

You MUST use this to write or revise developer-facing changelogs. Use for CHANGELOG.md files, unreleased sections, version entries, commit or pull-request categorization, breaking changes, deprecations, removals, fixes, and security changes.

You MUST use this to write or revise project README files. Use for repository, library, CLI, service, internal-tool, and open-source project introductions, installation, quick starts, usage, configuration, development, contribution, support, and license guidance.

Write or revise user-facing release notes. Use for product updates, version announcements, what's-new summaries, app-store notes, improvements, fixes, known issues, upgrade guidance, and action-required notices.