AI-SHIPR is not a single AI chat. It is a structured team of specialized agents, each holding a different part of your product context, equipped with skills they can run on demand.
Below is every component in the system — what it does, when to use it, and how it connects to your work.
You trigger. The system handles context. Memory compounds. Every component has a defined role — and they connect in one direction: toward sharper output over time.
Every decision logged, every hypothesis closed writes back to SHIPR files. The next session reads that context. Output gets sharper over time.
Agents hold a defined perspective and apply it consistently every time they are invoked. Unlike skills that run once, agents stay in context across a session.
Discovery & Framing
Forces clarity on product problems before solution work begins. Surfaces what outcome is actually needed, who is affected, and whether the framing links to strategy.
Decisions & Tradeoffs
Structures complex decisions by making options, criteria, and tradeoffs explicit. Prevents instinct-based choices from masquerading as analysis.
Alignment & Communication
Prepares alignment by reframing initiatives in stakeholder language. Adapts the message to who is in the room — exec, engineer, investor — without losing the substance.
Quality & Gaps
Audits SHIPR artifacts for structural gaps before they cause downstream damage. Checks initiative files, hypotheses, and strategy documents for completeness.
Research & Synthesis
Scans the Resources folder and surfaces what is relevant to current work. Turns accumulated reading into actionable product signals without you sorting it manually.
Interpersonal & Debrief
Handles interpersonal challenges, helps when stuck between options, and runs post-situation debriefs. The thinking partner for situations that are not purely analytical.
Initiative Strategy
Runs the full Lean Product Canvas (v3) in build, review, or export mode. Orchestrates Persona-Builder, Hypothesis-Builder, Experiment-Designer, and other skills — one box at a time or end-to-end.
Strategy & Direction
Builds complete product strategy using JTBD, Opportunity Solution Tree, and OKR. Operates one level above initiative work — defines what to build before the canvas determines how to validate it.
Lead Mode — Active when team_mode: lead in Settings.md
Portfolio & Allocation
Cross-product portfolio view — maps team initiatives to strategic bets, surfaces coverage gaps and overload risks, supports resource allocation decisions and exec portfolio updates.
People & Performance
Manages a team of PMs — prepares 1:1 agendas grounded in active initiatives, structures difficult feedback conversations, handles struggling PMs, and surfaces team health signals.
Skills are focused tools with one clear job. Invoke a skill when you need a specific output — not a conversation, not a session. Each skill knows what it needs from your strategy context before running.
JTBD Mapper — structured Jobs-to-be-Done analysis: job statement, functional/emotional/social dimensions, competing alternatives, unmet needs
Opportunity Tree — maps desired outcome to ranked customer opportunities, candidate solutions, and next riskiest assumption
Assumption Extractor — surfaces hidden assumptions in an initiative before you commit to it
Hypothesis Builder — converts an assumption into a falsifiable, testable statement
Research Synthesizer — converts raw research into structured product signals
Persona Builder — builds a structured user persona grounded in evidence
Edge Case Finder — stress-tests an initiative or experiment before it ships
Experiment Designer — designs the minimum viable experiment for the riskiest assumption: type, pass/fail criteria, structured card
Priority Stack — scores and ranks initiatives against each other using your current strategic bets
Tradeoff Mapper — compares two or more options structurally with explicit criteria
OKR Partner — sets metrics with specificity, bet linkage, and falsifiability
Narrative Refiner — adapts structured thinking for a specific stakeholder audience
Cross-Team Mapper — maps dependencies, handoffs, and owners across teams
1:1 Prep — builds a decision-first agenda for a manager or direct report 1:1
Board Update Builder — generates a structured product update for exec, board, or investor audiences
PRD Builder — generates a Product Requirements Document from a validated initiative
Build Companion — structures in-sprint PM decisions during active development
Build Review — prepares the sprint review comparing shipped vs committed scope
Fire Responder — classifies and structures a response to an active incident or unplanned work
Performance Tracker — validates or invalidates the hypothesis using post-launch data
Retro Facilitator — structures the sprint retro into actionable process improvements
Sprint Summarizer — generates sprint summary in internal or stakeholder mode
Business Case Builder — structures the initiative pitch with quantified impact and risk
Pitch Deck Builder — converts business case into a leadership presentation narrative
Create PM Voice — generates PM-Voice.md from your writing samples to calibrate output style
Workflows chain agents and skills together into structured multi-step processes. Run them on a cadence. The system keeps the structure so you can focus on the thinking.
Surfaces active initiatives, flags what needs attention, and gives you a prioritized start. Run every morning before you open a chat.
Maps active initiatives to sprint capacity. Outputs a sprint goal with committed items and explicit out-of-scope decisions. Run at the start of each sprint.
Strategic planning session. Reviews bet coverage, surfaces gaps, and sequences initiatives across time. Run during planning cycles.
Weekly structured review of all active initiatives. What is moving. What is stuck. What needs a decision this week. Run every Monday or start of week.
Weekly review of all active and completed experiments. Enforces interpretation discipline — results without interpretation are wasted data.
Pre-launch go/no-go checklist for any initiative. Surfaces open decisions and risks before the release. Run 3 to 5 days before a planned launch.
Closes one cycle and opens the next based on what data says. Run after a measurement window completes or a hypothesis is validated or invalidated.
Close the week with a structured review of what moved, what was decided, and what was learned. Updates Learning.md. Run every Friday.
Initial setup workflow. Runs once after you fill in the Setup Worksheet. Populates Strategy files from your raw input. Not a daily command.
Playbooks are pre-written prompts for specific situations. You do not need to know which agents or skills to invoke. Paste the right playbook into Claude Code and the system takes over — loading the correct intelligence layer for your situation.
A stakeholder drops an idea. The playbook turns it into a problem frame, assumption map, risk score, and falsifiable hypothesis — before any build decision is made.
Beginning a new sprint with full product context. Surfaces active initiatives, validates scope, and outputs a sprint goal with committed items and explicit tradeoffs.
3 to 5 days before a planned release. Coordinates the last mile — go/no-go checklist, open decisions surfaced, launch criteria reviewed.
Post-launch metric read. Validates or invalidates the hypothesis, identifies what to do next, and logs the learning before it gets buried in notes.
Preparing a leadership or funding pitch. Builds the business case, structures the narrative, and adapts the message to the specific stakeholder in the room.
Active incident or production issue. Classifies severity, assigns response structure, and keeps decision-making clear when things are moving fast.
Morning before a manager or direct report 1:1. Reads your initiatives, surfaces open decisions, and builds a structured agenda — decisions first, not status first.
Friday loop close. Reviews what moved, logs decisions made, captures learnings, and frames priorities for next week. Updates Learning.md so the system compounds.
Adding a new persistent AI persona to your AI-SHIPR system. Guides the definition of role, scope, triggers, and required reading for the new agent.
Adding a new single-task skill to extend the system. Structures the skill definition, its trigger conditions, inputs needed, and expected output format.