mituso89

Audit a Jira or Azure DevOps board against ways-of-working rules, produce a tagged cleanup plan, then on one confirm gate apply safe bulk fixes and delegate epic/story/subtask creation to sibling skills. Triggers: audit my board, board audit, ado board audit, jira audit, audit board, backlog cleanup, board hygiene, board health, clean up jira, sprint cleanup, iteration cleanup, sprints used as epics, hidden tickets, fix backlog, organize backlog, epic story audit, backlog triage.

Shared board-provider layer — detect whether this session talks to Jira or Azure DevOps (ADO), load the matching provider mapping (tool names, query language, types, states), and record the choice in the calling skill's config.json. Board-facing skills delegate their provider detection here and stay platform-neutral. Triggers: which board provider, board setup, connect jira, connect ado, azure devops setup, board provider.

Author a new Claude Code skill — folder + SKILL.md + frontmatter, lean 50–120 lines, grounded in the host project's real files and conventions, integrated with sibling skills. Triggers: create skill, new skill, add skill, write skill, skill for, author skill, scaffold skill.

Standalone task orchestrator — take a request (or a board ticket) from GO gate through implement → test → verify → review → finalize by delegating to sibling m-skills. Works with or without an SDD spec. Triggers: implement, build, create feature, add feature, new screen, new page, task, scaffold, build X, ship this ticket, work on this ticket, do the ticket.

PM ticket authoring at the right altitude — turn a bug report, feature ask, or an over-technical draft into ONE clean ticket/work item that stops at Problem, Impact, Reproduction, and Acceptance Criteria, stripping engineering diagnosis (root cause, code, file:line, commit hashes, proposed fix) and leaving it to engineering. Works against Jira or Azure DevOps via the m-board provider layer; creates the ticket only after a confirm gate. Triggers: write a ticket, pm ticket, create work item, ado ticket, jira ticket, bug ticket at the right altitude, de-technify this ticket, file a bug, report a bug, raise a ticket, bug report, feature request, my ticket is too technical, clean up this ticket, acceptance criteria, problem impact reproduction, ticket too engineering-heavy.

QA deliverable generator — produce test plans, manual test cases, regression suites, design-vs-implementation validations, and structured bug reports from a feature description, in Markdown ready for any tracker. Optional Figma MCP integration for pixel-level design validation. Triggers: test plan, test cases, regression suite, QA, bug report, figma validation, manual testing, smoke test, test coverage.

Meta audit of a SKILL.md — frontmatter shape, trigger coverage, grounding to real files, compliance with the HOST project's own conventions, integration with sibling skills, length & tone. Triggers: review skill, audit skill, check skill, skill quality, validate skill, skill review.

SDD phase 7 — the SDD adapter around the m-implement engine: gate on tasks+analyze, human GO, then delegate the implement→test→verify→review→finalize loop to m-implement, and record/route the SDD state. Triggers: sdd implement, execute sdd plan, build the feature sdd, run sdd tasks, implement sdd.

Atomically update all 6 Fluent learner databases (learner-profile, progress, mistakes, mastery, spaced-repetition, session-log) at session end by calling .claude/hooks/update-db.py with a single JSON payload. Use at the end of every practice session — fluent-writing, fluent-vocab, fluent-speaking, fluent-reading, fluent-review, fluent-learn — to persist the session's errors, review results, new vocabulary, and session metadata.

List, switch, and add languages for multi-language learners — each language gets its own isolated set of the 6 databases under data/<slug>/, tracked in data/languages.json. Triggered only when the learner types /fluent-language (list / switch <language> / add). Never auto-invoke — switching or adding changes which database directory every other skill reads from next.

Main adaptive language-learning session that mixes skills (writing, speaking, vocabulary, reading) and exercise types based on the learner's current level, weak patterns, and due reviews. Triggered only when the learner types /fluent-learn. Greets the learner, shows today's plan, asks what to practice, runs interleaved exercises one at a time, and updates all databases at the end.

Run an interactive reading comprehension session with a short target-language text followed by main-idea, detail, vocabulary-in-context, inference, and true/false questions. Triggered only when the learner types /fluent-reading. Presents the text, waits for the learner to read, then asks questions one at a time with immediate feedback, and optionally adds new vocabulary to the spaced-repetition queue.

Run today's spaced-repetition review queue — items scheduled by SM-2 that need reinforcement before the learner forgets them. Triggered only when the learner types /fluent-review. Pulls due items from spaced-repetition.review_queue.today, generates a targeted exercise for each, evaluates the response, updates SM-2 parameters, and reshelves items into the correct future queue.

Parse Fluent `/results/*.md` session files to extract error patterns, strengths, accuracy trends, and focus areas for the next session. Use when the tutor needs to analyze the learner's recent performance — planning the next lesson, recommending focus areas, or answering "what should I practice next?".

One-time interactive onboarding that creates the learner's personalized language-learning profile — name, target language, native language, current/target CEFR level, timeline, daily minutes, and learning goals. Triggered only when the learner types /fluent-setup. Also handles profile updates and resets for returning users. Must never auto-invoke because re-running can reset progress.

Generate a voice speaking session card (default 10 questions, configurable via preferences.speaking_export_question_count) and hand it off to the Claude app's voice mode — via a Slack channel when the claude.ai Slack connector is available, Google Drive as fallback, otherwise as a local file to paste. Triggered only when the learner types /fluent-speaking-export. Questions mix level topics, the learner's top mistake patterns, due review vocabulary, and one role-play. After the voice session, /fluent-speaking-import brings the results back.