Sprint writing,
without the fluff.

May 5, 2026 · 9 min read One Rails app, two dynos — running a public MCP server alongside the main SaaS Engineering-honest writeup of how SprintFlint runs both the user-facing web app and the public MCP server (mcp.sprintflint.com) from a single Rails 8 codebase, deployed twice on Heroku via the multi-procfile buildpack. Why two dynos beat one (blast-radius isolation), why one repo beats two (no schema drift, shared auth, coordinated deploys), three things we got wrong on the first attempt (release running twice, log levels, MCP routes in main routes.rb), and when this shape starts to creak (different scaling profiles, different security posture, different language fits). May 5, 2026 · 9 min read Claude Desktop for sprint management — five real use cases (and three things it's not good at) Claude Desktop with MCP turns your sprint board into something you can ask questions of. Five use cases that justify the setup (Monday triage, what-to-work-on-next, bug-report-to-ticket, status updates, retro prep), three things it's bad at (strategy, destructive ops, long-horizon forecasting), and the practical guide to setup, tokens, and what NOT to expose to the agent. May 5, 2026 · 8 min read Sprint state in your agent — what changes when MCP makes the board reachable from the editor Most AI coding agents don't know which sprint you're in. MCP fixes that. This is the practitioner's version: what the protocol is, the four real-world patterns it unlocks (pull-mode planning, auto-PR comments, sprint-aware review, standup pre-write), the eight tools we shipped on day one (and the two we deliberately didn't — both delete operations), and the four ways teams misuse the integration in the first two weeks. May 5, 2026 · 13 min read Running sprints with AI teammates — what actually changes (planning, velocity, code review, retros) When part of your team is Cursor / Claude Code / Devin / Aider agents that pick tickets, ship code, and close PRs, your sprint cycle stops fitting the shape it had. Here's the practical version: what changes in planning (two queues, spec-quality gate), velocity (it goes up and gets noisier), code review (the new bottleneck), retros (the agent setup is half your throughput now), and the four mistakes teams make in the first three sprints. May 2, 2026 · 12 min read Agile estimation techniques compared — story points, planning poker, t-shirts, hours, no-estimates Five real estimation techniques, what each is good at, what each breaks under, and how to switch without burning a sprint. Story points, planning poker, t-shirt sizing, ideal hours, no-estimates — comparison table, when each fits, the three questions to ask before switching, and the one rule that matters more than the technique. Apr 29, 2026 · 8 min read 5 sprint workflow toolkits — what to use when (planning, stories, forecasting, retros, metrics) Sprint workflow has five distinct stages, each with its own failure mode. This is the hub: five free toolkits mapped to where sprints actually break — planning, story management, forecasting, retrospectives, metrics. Triage by symptom, pick the toolkit, fix the structural problem. All free, no signup. Apr 26, 2026 · 10 min read Forecasting sprint dates with p85 (not the average): the math and the message Most sprint forecasts miss because they're built on the mean. Cycle times have long tails — the average tells you nothing about whether this ticket will take 3 days or 12. Switch to p85, build the tail into the budget, and dates start landing. Here's the math, the stakeholder message, and the four anti-patterns teams hit when they switch. Apr 23, 2026 · 12 min read Tech debt: when to pay it, when to ignore it (with the calculation) Most teams handle tech debt with permanent guilt and a tech-debt sprint that never quite happens. Here's the practical version — three categories (compounding, static, frozen) that need different handling, the cost calculation that decides whether to pay, five anti-patterns that waste tech-debt budget, and how much of a sprint should be debt work (10-25%, with the why). Apr 20, 2026 · 11 min read Definition of Ready: 5 entry criteria that prevent half-baked sprints Most teams have a Definition of Done. Few have a Definition of Ready — and that's why stories enter sprints in wildly different states. Five entry criteria that catch most surprises (clear why, testable acceptance criteria, estimable size, dependencies identified, owner assignable), what to refuse adding, how DoR fits with refinement and planning, and how to introduce it without turning refinement into a gatekeeping ceremony. Apr 17, 2026 · 11 min read Acceptance criteria: 5 formats and the structural traps that ruin them Acceptance criteria are the most-skipped part of the most-skipped artefact. Here's the practical guide — five formats (bullets, Gherkin, examples table, test list, DoD-plus-delta), when each one fits, five anti-patterns that wreck criteria regardless of format, the five-minute rule and the three-bullet minimum for teams starting from zero. Apr 14, 2026 · 11 min read Working with non-engineering stakeholders: the engineering manager's translator playbook Stakeholders aren't enemies — they're the source of the work, and they need a competent translator. The translation problem (same words, different meanings), five recurring stakeholder personalities (eager PM, anxious founder, skeptical finance, domain expert, disengaged) with structural fixes for each, the weekly update template, and the trade-off-visibility rule that protects the team without burning the relationship. Apr 11, 2026 · 12 min read What makes a good user story (INVEST applied, with side-by-side examples) Most user stories are written because the template requires it, then ignored. Here's the practical version — when the role template helps and when it lies, INVEST applied with concrete tests, three side-by-side bad-vs-good examples (feature, infrastructure, bug), the 30-second rule that fixes 80% of ticket quality, and five anti-patterns to refuse on sight. Apr 8, 2026 · 11 min read Engineering manager burnout: 6 sprint patterns that signal it (and the structural fixes) EM burnout shows up as sprint patterns long before the EM admits it. Six observable signals — Sunday-merge burndown, escalations that never trend down, no shipped code in 8+ weeks, sprint goals that quietly disappear, 70% meeting calendars, EM stops enjoying the work — with structural fixes for each. For engineering managers, the people who manage them, and engineers who want to help. Apr 5, 2026 · 10 min read Mid-sprint scope changes: when to absorb them, when to defer (with the script) Stakeholder messages day 4 of a 10-day sprint asking for one more thing. Here's the framework — three questions to ask before saying yes, four scenarios with explicit playbooks (production incident, urgent escalation, 'quick win', new information), why you need a 15-30% sprint reserve, and the exact script to say no without burning the relationship. Apr 2, 2026 · 9 min read Sprint backlog vs product backlog (the actual difference + 4 ways teams blur them) Confusing sprint and product backlog causes most planning chaos on agile teams. Here's the honest read on the difference, four blurring patterns (wishlist sprint, sprint queue in backlog, mid-sprint dumps, refinement-at-planning), the structural fix for each, and a simpler model for very small teams. Mar 30, 2026 · 11 min read Velocity dropped — here's the actual playbook (don't panic, don't pad) Velocity dropped 30% over two sprints? Here's the calm playbook — when a dip is noise vs signal, the seven real causes (in order of probability), the one-meeting planning fix, and what NOT to do under pressure (more ceremonies, 'commit harder', resetting story points). Mar 27, 2026 · 9 min read Sprint review without the demo theatre (what it's actually for) Most sprint reviews are demo theatre — engineers parade work, stakeholders nod, nothing changes. Here's the honest read on what sprint review is actually for, three signals you've drifted, the fix (open with goal, demo with intent, reserve 20 minutes for 'now what'), and a 60-minute agenda that works. Mar 24, 2026 · 10 min read The agile metrics that actually help (and the ones to ignore) Most agile dashboards measure typing volume, not shipping. Here's the honest read — the four metrics that pay off (velocity, sprint goal hit rate, cycle time, escape rate), the five that don't (LOC, hours, points-per-dev, burndown shape, ticket count), and how to use them without breaking the team. Mar 21, 2026 · 9 min read The hidden cost of your daily standup (run the math) Daily standup costs more than your AWS bill. Here's the math — £40,000-£68,000/year for a 6-person team — three diagnostic signs you're overpaying, and four standup formats from cheapest (async written) to most expensive (classic three-question), with what each actually buys. Mar 18, 2026 · 8 min read Backlog refinement: the 30-minute runbook (and what should never happen in it) Most teams half-skip refinement and pay for it in planning. Here's the practical runbook — what refinement actually produces, the 30-minute weekly version, two anti-patterns that ruin it (silent dump, open-ended discussion), and what should never happen in the meeting. Mar 15, 2026 · 10 min read How to estimate story points: the only practical guide you'll actually use Most story points content is vague training-deck filler. Here's the practical version — what story points actually measure, the modified-Fibonacci scale that works, the baseline-setting step most teams skip, the 30-seconds-per-ticket loop, six patterns for splitting too-big stories, and what AI estimates are worth. Mar 12, 2026 · 9 min read Kanban vs scrum: how to actually pick (and the hybrid most teams run) Scrum is a cadence; kanban is a flow. The honest read on which fits your team — the four signals that pick for you, when scrum breaks down, when kanban breaks down, and the scrumban hybrid most successful product teams actually run. Mar 9, 2026 · 9 min read Agile vs scrum: the actual difference (and why it matters for your team) Most teams use 'agile' and 'scrum' interchangeably. They're not the same. Here's the honest read — what each one is, how teams do scrum without being agile, when to pick each, and the practical impact on tool choice. Mar 6, 2026 · 9 min read Burndown charts: when they help, when they lie, and how to actually read one Burndowns are the most photogenic agile artefact and the most misread. Here's the honest read — when they tell the truth, the four ways they lie, what to pair them with, and the playbook for actually using one. Mar 3, 2026 · 9 min read The 60-minute sprint planning agenda (and the 5 things to skip) Most planning meetings run 2 hours and produce uncertain commits. The fix is agenda design, not effort. Here's a 60-minute agenda for a 2-week sprint, the prep that makes it work, and 5 things to drop from the standard playbook. Feb 28, 2026 · 8 min read What 'definition of done' should actually mean for engineering teams The 12-bullet DoD on every team's wiki is decoration. Here's the engineering version that catches real bugs — 6 verifiable items, what to drop from the standard list, and how to roll it out without ceremony. Feb 25, 2026 · 8 min read How a 5-person engineering team should actually run sprints (and what to skip) Most agile content is written for 30-person orgs. Here's the practical version for small engineering teams — the minimum stack of 5 rituals, the 7 ceremonies you should explicitly skip, and the cadence that consistently works. Feb 19, 2026 · 8 min read Why most retrospectives die after 3 sprints (and how to keep them useful) Retros don't fail because teams stop caring. They fail for seven structural reasons — actions that never ship, formats that never change, and a lead who does most of the talking. Here's how to keep them alive. Feb 16, 2026 · 7 min read What a sprint goal should actually sound like (12 real examples) 'Ship the new dashboard' isn't a sprint goal. Here's what good ones look like, what bad ones sound like, and 12 real examples by team type — engineering, platform, mobile, security, design. Feb 13, 2026 · 9 min read Sprint Velocity Benchmarks: What numbers healthy teams actually hit What's a 'good' sprint velocity? Honest ranges by team size and sprint length, the five variables that move the number, and the diagnostic questions that beat any benchmark.