risadams/ink-and-agency

Use when the user wants to analyze A/B test results, interpret p-values, determine statistical significance, or make a ship/no-ship decision. Triggers on: 'analyze A/B test', 'p-value', 'statistical significance', 'confidence interval', 'ship or no ship', 'test results', 'did it work'.

Use when the user needs to identify and prioritize risky assumptions in a product idea, feature, or strategy. Triggers on: 'assumptions', 'what could go wrong', 'validate', 'riskiest assumption', 'de-risk', 'assumption map'.

Use when the user needs to groom, refine, or clean up a product backlog. Triggers on: 'groom backlog', 'backlog refinement', 'backlog grooming', 'clean up backlog', 'refine stories', 'sprint refinement', 'backlog management'.

Safely rebase the current branch onto its upstream target, auto-resolving trivial conflicts and prompting for complex ones. Use when the user says "rebase", "rebase onto main", "update my branch", "sync with baseline", "catch up with main", or invokes /branch- rebase.

Use when a git merge or rebase hits complex conflicts, or the user says "resolve conflicts" / "fix merge conflicts". Reconstructs intent, preserves both branches' goals when possible, and runs automated checks.

Decode messages into plain language with tone and intent analysis. Breaks down what someone actually means, analyzes emotional tone with probability scores, and flags ambiguity. Use when user says "break it down", "what does this mean", "decode this", "analyze this message", "what are they saying", or invokes /break-it-down.

Facilitates structured single-persona or multi-persona consultation for strategy, product, engineering, operations, and risk decisions. Use when the user requests a council, multiple perspectives, persona advice, tradeoff analysis, or iterative clarification and debate.

Two-axis review of the working diff since a fixed point — Standards (repo conventions and code smells) and Spec (matches the originating ticket). Use when the user wants to review a branch or work-in-progress changes, says "review since X", or asks for a pre-MR review.

Analyze a git repository's history and produce an SVG treemap of file churn — area = lines changed, color = commit frequency. Use when user wants to find unstable areas of a codebase, predict bug-prone files, see "where do we change the most", spot refactor candidates, or invokes /codebase-churn.

Shared vocabulary and principles for designing deep modules — a lot of behaviour behind a small interface, placed at a clean seam. Use when designing or improving a module's interface, deciding where a seam goes, making code more testable or AI-navigable, or when another skill needs the deep-module vocabulary.

Tell the agent to explain code with broader context and a higher-level perspective. Use when you're unfamiliar with a section of code or need to understand how it fits into the bigger picture.

Find deepening opportunities in a codebase, informed by the domain language in CONTEXT.md and the decisions in docs/adr/. Use when the user wants to improve architecture, find refactoring opportunities, consolidate tightly-coupled modules, or make a codebase more testable and AI-navigable.

Create a detailed refactor plan with tiny commits via user interview, then file it as an issue. Use when user wants to plan a refactor, create a refactoring RFC, or break a refactor into safe incremental steps.

Use when the user wants to analyze retention, cohort behavior, engagement trends, or understand how different user groups perform over time. Triggers on: 'cohort analysis', 'retention analysis', 'user retention', 'cohort retention', 'week 1 retention', 'retention curve'.

Personal daily briefing assistant. Pulls the last 24 hours of email and calendar from Outlook, summarizes today's schedule, surfaces open action items, and proposes focus blocks. Use when user says "daily briefing", "morning briefing", "daily standup prep", or invokes /daily-briefing.

Generate a per-team standup report from recent Jira, GitLab, Confluence, and Git activity, written into the Obsidian vault. Use when the user says "daily standup prep", "standup report", "generate standup", "scrum prep", or invokes /daily-standup-prep.

Diagnosis loop for hard bugs and performance regressions. Use when the user says "diagnose"/"debug this", or reports something broken/throwing/failing/slow.

Extract clean markdown from web pages using Defuddle CLI, stripping navigation and clutter to save tokens. Use instead of WebFetch for any standard web page, documentation, article, or blog post — but not for URLs ending in .md, which WebFetch handles directly.

Actively build and sharpen a project's domain model — challenge fuzzy terms, stress-test with scenarios, and write CONTEXT.md and ADRs inline as decisions crystallise. Use when the user wants to pin down domain terminology or a ubiquitous language, record an architectural decision, or when another skill needs to maintain the domain model.

Score today's load (meetings, sensory demand, social events, deep work), flag burnout risk, and suggest what to drop or defer. Use when the user says "energy budget", "spoon check", "am I overcommitted", "what should I drop", or invokes /energy-budget.

Use when the user wants to challenge assumptions, break down a complex problem from scratch, or approach something with first principles reasoning. Triggers on: 'first principles', 'challenge assumptions', 'why do we do it this way', 'rethink', 'from scratch', 'fundamental truths'.

Morning kickoff wrapper running sprint-snapshot, daily-standup-prep, and daily-briefing in order; idempotent on same-day re-runs. Use when the user says "good morning", "morning routine", "kick off my day", "start my day", or invokes /good-morning.

Interview the user relentlessly about a plan or design until reaching shared understanding, resolving each branch of the decision tree. Use when user wants to stress-test a plan, get grilled on their design, or mentions "grill me".

Grilling session that challenges your plan against the existing domain model, sharpens terminology, and updates documentation (CONTEXT.md, ADRs) inline as decisions crystallise. Use when user wants to stress-test a plan against their project's language and documented decisions.

Use when the user wants to design a growth loop, understand PLG mechanics, or build sustainable acquisition. Triggers on: 'growth loop', 'flywheel', 'viral loop', 'PLG growth', 'product-led growth', 'growth mechanics', 'how do we grow', 'word of mouth'.

Compact the current conversation into a handoff document for another agent to pick up.

Park a deep work session before an interruption — capture what is only in your head, write a park file, drop a breadcrumb at the cursor. Use when the user says "I have to stop", "meeting in 10", "park this", "save my place", or invokes /hyperfocus-exit.

Reconstruct context after a deep session or interruption from git diff, modified files, and recent commits, then propose the next concrete action. Use when the user says "what was I doing", "where did I leave off", "lost my place", or invokes /hyperfocus-recovery.

Shape every response for an autistic reader — direct statements, required/optional labels, literal language, quantified uncertainty, named social subtext. Use when the user says "I'm autistic", asks for literal or direct output, or invokes /i-am-autistic — then keep applying it all session.

Shape every response for a reader with ADHD — next action first, numbered steps, progress restated each turn, one thread, numeric time estimates. Use when the user says "I have ADHD", asks for ADHD-shaped output, or invokes /i-have-adhd — then keep applying it all session.

Weigh one or more ideas with a structured pros-and-cons analysis and a Go / Go with caveats / Don't recommendation. Use when the user wants to evaluate an idea before committing, compare options on the merits, score tradeoffs, or asks "should I do X" / "is X worth it".

Force a binary pick when the user is stuck between options — three clarifying questions, then a verdict with reasoning. The fast counterpart to idea-choice. Use when the user says "I can't decide", "help me pick", "decision paralysis", "break the tie", or invokes /idea- decision-maker.

Helps users generate, refine, and stress-test ideas from loose topics. Use when user asks to brainstorm, find new ideas, or needs creative inspiration for a topic.

Build a piece of work (ticket, spec, or free-form request) through a tunable pipeline — test-first, verify, review, and track — where every phase can be skipped or tuned. Use when the user wants to actually build something, from a throwaway prototype to a CCB-gated feature.

Capture the current hyperfixation into the Obsidian vault fast, so it isn't lost and doesn't derail today's work. Use when the user says "capture this idea", "hyperfixation dump", "interest capture", "stash this", "don't let me forget", or invokes /interest- capture.

Draft a customer-facing release note for a Jira ticket from its context and linked merge requests. Read-only unless the user opts in to publish. Use when the user says "draft release notes", "write a release note", "release note for [TICKET-KEY]", or invokes /issue-draft-release-notes.

Scan a Jira project's open issues for likely duplicates and report each suspected pair with a probability score and reasoning. Read-only. Use when the user says "find duplicates", "duplicate check", "dedupe issues", or invokes /issue-dup-find.

Estimate story points for a Jira ticket using historical data and multi-persona scrum poker. Use when user says "estimate this ticket", "how many points", "story point estimate", "estimate [TICKET-KEY]", or invokes /issue-estimate-sp. Read-only — never modifies the ticket.

Analyze a Jira ticket into an actionable breakdown, following linked tickets and Confluence context to surface gaps and ambiguities. Read-only. Use when the user says "break down this feature", "analyze this ticket", "feature breakdown", or provides a ticket key for analysis.

Suggest Jira components for one ticket or sweep a project's open tickets, confirming before any modification. Use when the user says "suggest components", "tag components", "fix components", "categorize tickets", or invokes /issue-suggest-component.

Triage a bug or Jira ticket into ranked root-cause candidates and proposed solution paths. Read-only unless the user opts in to publish. Use when the user says "triage this", "triage [TICKET-KEY]", "what's causing this bug", "help me find the root cause", or invokes /issue-triage.

Process a meeting dump — separate facts from feelings, flag action items, sort social ambiguities into worth-following-up vs noise. Use when the user says "meeting decompression", "process this meeting", "what just happened", or invokes /meeting- decompression.

Create and edit Obsidian Bases (.base files) with views, filters, formulas, and summaries. Use when working with .base files, creating database-like views of notes, or when the user mentions Bases, table views, card views, filters, or formulas in Obsidian.

Create and edit Obsidian Canvas files (.canvas) following the JSON Canvas spec — nodes, edges, groups, and connections. Use when working with .canvas files, creating visual canvases, mind maps, or flowcharts in Obsidian.

Build interactive Chart.js charts in Obsidian notes via chart codeblocks — inline series, a markdown table by block ID, or a Dataview query. Use when the user wants to chart, plot, graph, or visualize data in their Obsidian notes, or mentions "obsidian charts" or "renderChart".

Read, create, search, and manage Obsidian vault notes, tasks, and properties from the command line; also supports plugin and theme development and debugging. Use when the user asks to interact with their Obsidian vault or develop Obsidian plugins and themes.

Create and edit Obsidian Flavored Markdown with wikilinks, embeds, callouts, properties, and other Obsidian-specific syntax. Use when working with .md files in Obsidian, or when the user mentions wikilinks, callouts, frontmatter, tags, embeds, or Obsidian notes.

Search, create, and manage notes in the Obsidian vault with wikilinks and index notes. Use when user wants to find, create, or organize notes in Obsidian.

Turn an unstructured dump of everything competing for attention into exactly one thing to do now, with the rest parked and dated. Use when the user says "everything is on fire", "I'm overwhelmed", "too much going on", "I don't know what to do first", or invokes /overwhelm-triage.

Turn a decision you can't answer alone into a Markdown questionnaire for the one person who can — filled in async, or worked through together in a meeting.

Synthesize the current conversation into a spec (PRD) and stage it as a local file. Use when a design discussion, grilling session, or work-plan is settled and you want it written up as a spec — no interview, just synthesis. Publishing to Confluence is opt-in, only on request.

Break a spec, plan, or the current conversation into a local, numbered, spec-kit-style task list of tracer-bullet vertical slices with explicit dependencies. Use when a spec or plan is ready to become actionable tasks. Pushing to Jira is opt-in, only on request.

Build a throwaway prototype to answer a design question. Use when the user wants to sanity-check whether a state model or logic feels right, explore what a UI should look like, or resolve a prototype-type decision ticket before committing to a build.

Write, overhaul, or review a project README that is accurate to the repository and readable by a human. Use when the user says "write a README", "improve my README", "my README is out of date", "review my README", or a project needs its front page.

Calibrate a message that stung — separate evidence from interpretation, score how much sting is warranted. Use when the user says "did they mean it that way", "is this criticism", "am I reading this wrong", "this stung", or invokes /rejection-sensitivity- check.

Investigate a question against high-trust primary sources and capture the findings as a Markdown file in the repo. Use when the user wants a topic researched, docs or API facts gathered, or reading legwork delegated to a background agent.

Create new agent skills with proper structure, progressive disclosure, and bundled resources. Use when user wants to create, write, or build a new skill.

Convert the start-of-sprint canvas into a planning report — committed scope, capacity vs commitment, WIP-saturation risk, observations, risks. Use when the user says "sprint plan", "sprint planning report", or invokes /sprint-plan.

End-of-sprint stakeholder report comparing the sprint's start and end canvases, fitted to the standard Scrum Master template. Use when the user says "sprint review", "end of sprint report", "sprint stakeholder report", or invokes /sprint-review.

Capture a point-in-time snapshot of a scrum team's sprint board into the Obsidian vault. Use when the user says "sprint snapshot", "snapshot the sprint", "capture sprint state", "sprint planner", "scrum board snapshot", or invokes /sprint-snapshot.

Weekly scrum-of-scrums report comparing two sprint snapshots — findings, observations, trouble areas, trends. Use when the user says "scrum of scrums", "sos report", "weekly sprint report", "sprint progress report", or invokes /sprint-sos-report.

Produce the literal first physical action when the user knows what to do but can't start — not a plan, not a breakdown. Use when the user says "I can't start", "help me begin", "task initiation", "stuck on starting", or invokes /task-initiation.

Test-driven development — the red → green loop and the rules that make its tests worth keeping. Use when the user wants to build a feature or fix a bug test-first, mentions "red-green-refactor", TDD, or integration tests.

Teach the user a new skill or concept, within this workspace.

Counter time blindness with a calibrated estimate that accounts for setup, interruptions, and re-entry. Use when the user says "how long will this take", "time check", "reality check", "estimate this", "am I being realistic", or invokes /time-reality-check.

Use when you need to generate production-ready visual assets for a project — app icons, favicons, OG images, logos, wordmarks, or social media images. Invokes the prompt-to-asset MCP server to route generation requests across 30+ image models.

Stop — that last message didn't land. Re-pitch it in plain language with the context I'm missing.

Router over this pack's skills and flows. Ask it "which skill fits X" when you don't remember what's available.

Generate an interactive bash wizard that walks a human step by step through a manual procedure. Use when provisioning infrastructure, setting up credentials or CI secrets, walking someone through an unfamiliar third-party dashboard, or running a one-off migration or cutover. Don't invoke it for steps the agent can perform itself.

Plan a huge chunk of work — more than one agent session can hold — as a shared map of decision tickets on your issue tracker, and resolve them one at a time until the way to the destination is clear.

Calibrate a drafted apology — strip reflexive over-apology, keep warranted accountability. Use when the user says "calibrate this apology", "am I over-apologizing", "apology check", "did I say sorry too much", or invokes /writing-apology-calibrator.

Shape an article as a journey of beats, choose-your-own-adventure style — write one beat, offer pivots, repeat until it ends. Use when the user has raw material and wants it assembled as a narrative rather than an argument.

Produce the first sentence of a message when the blank cursor is winning — two or three openings matched to relationship and goal, then step out. Use when the user says "cold open", "help me start this message", "how do I begin", or invokes /writing-cold-open.

Guide the user from raw idea or draft to a finished article through relentless questioning and iterative writing. Use when the user wants to write something from scratch and have it completed.

Reference for writing any document an agent consumes — a skill, an AGENTS.md or CLAUDE.md, a doc reached by a pointer. Use when creating or editing a skill, writing agent-facing instructions, or modifying AGENTS.md / CLAUDE.md.

Mine the user for raw writing fragments — claims, vignettes, sharp sentences, half- thoughts — and append them to one document as material for a future article. Use when the user wants to develop ideas before imposing structure, or mentions "fragments" or "raw material".

Remove signs of AI-generated writing — inflated symbolism, promotional language, vague attributions, em dash overuse, rule of three, AI vocabulary, negative parallelisms, filler. Use when editing or reviewing text to make it read as human-written.

Shape a file of raw material into an article through a conversational session — candidate openings, paragraph by paragraph, arguing format at each step. Use when the user has notes, fragments, or a rough draft and wants it turned into something publishable.

Generate a script for a dreaded social scenario — declining, asking for an extension, following up, setting a boundary — in two or three phrasings from direct to softened. Use when the user says "script for", "help me say", "how do I tell them", or invokes /writing- social-script.

Reverse decoder — paste a draft message and get a read on how it will land before you send. Flags blunt, cold, accidentally passive-aggressive, or over-apologetic patterns. Use when user says "tone check", "how does this sound", "before I send this", "will this land okay", "check this draft", or invokes /writing-tone-check.