# Strategic Realignment > When the operator says "what's next?" / "review the plans" / "where are we?" — the question is rarely "list our open tickets." It's "help me see the shape of the next arc." The skill is to cross-ground reality (recent commits, live state, observed metrics) against the canonical strategy docs, surface the genuine drift, present 2-3 named options with tradeoffs, and let the operator decide. Don't decide for them; don't list everything; don't restate the strategy. Use when this capability is needed. - Skill: `tomevault-io/strategic-realignment` (Agent Skill, multi-file: 2 files) - Install (CLI): `npx skillmds@latest add tomevault-io/strategic-realignment` - Raw SKILL.md: https://api.skillmd.com/api/skills/tomevault-io/strategic-realignment/raw - Safety review: pending (external: skill-scanner PASS, skillspector PASS) - Works with: Claude Code, Claude.ai, OpenAI Codex - Category: Coding & Dev Tools - Author: tomevault-io (https://skillmd.com/u/tomevault-io) - Updated: 2026-09-17 - Page: https://skillmd.com/skills/tomevault-io/strategic-realignment --- # Strategic realignment The operator asking "what's next?" is doing strategic thinking, not asking for a queue dump. They have a felt sense that the current direction needs revisiting. The skill is to translate that felt sense into a concrete reframe they can react to — not to enumerate work, not to decide for them, not to deliver a strategy lecture. This is the highest-leverage conversational pattern in long-running projects. Done well, a 20-minute realignment redirects weeks of build. Done badly, it pads the operator's inbox with ten options they didn't want. ## When to apply The operator initiates with phrasings like: - "What's next?" - "Let's review the plans." - "Where are we?" - "I feel like we should step back." - "Is this still the right priority?" - "Have we drifted from X?" - "Help me see what we should be doing." Or implicitly when: - The operator's recent decisions show they're scope-questioning, not just executing. - A meaningful event happened (incident resolved, gate signal moved, competitor moved, deadline shifted) and no one's reframed yet. - The work has been heads-down for >2 weeks without a step-back. Skip the ceremony when: - The question is narrow ("what should I do *today*?"). Answer narrow. - The operator already named the reframe and is asking you to cost it. - A realignment happened in the last 7 days and nothing material has changed. ## The sequence 1. **Read the canonical strategy docs first.** Whatever your project's authoritative strategy file is — roadmap, vision doc, OKRs, north-star memo — load it before generating any options. Strategy lives there or nowhere; do not ad-lib direction from session context. 2. **Cross-ground against live state.** What does the strategy file say is current? What does reality say? Likely candidates for drift: - Stage/phase says X is current; build state says X shipped two weeks ago. - Strategy says we measure metric Y; metric Y is broken or missing. - A gate signal silently flipped (positive or negative) and no one noticed. - A risk on the watch-list became real and isn't reflected in the priorities. 3. **Read the recent operator-shaped artefacts.** Last session debriefs, last 30 commits, last decisions, recent feedback. The operator's thinking has been laid down in those — surface it. 4. **Identify the binding constraint.** What's actually limiting forward motion right now? Common shapes: - Volume — we need more of X (input, output, signal). - Quality — we have enough X but the quality is wrong. - Distribution — we have it but no one sees it. - Demand — supply is healthy, demand is cold. - Trust — the system works but readers don't believe it. - Attention — the work is right but the founder's time is mis-allocated. The binding constraint is rarely the same as the loudest alert. 5. **Present 2-3 named lanes, not a checklist.** Each lane needs: - A name that fits the strategic frame, not the implementation detail. - One sentence on why it matters now. - The honest tradeoff against the alternatives. - One concrete first move. Naming is load-bearing. "Lane 2 — lead the category" is reactable; "do these 17 things" is not. 6. **Recommend, then ask.** Pick one lane and say why. Be willing to be wrong — the operator's role is to redirect. Frame: "I'd start with Lane X because ; Lane Y is the alternative if ." 7. **Don't decide.** The operator is doing strategic thinking. Your role is to surface the shape, present the tradeoffs, and recommend. The decision is theirs. Do not author plans, do not commit to a direction, do not start implementing — until the operator confirms which lane. ## What to NOT do - **Restate the strategy.** They wrote it; they don't need it back. Reference, don't recite. - **Dump the open ticket queue.** "What's next?" is not "what's pending?" Open tickets are the *output* of strategy, not the strategy. - **Generate ten options.** The cost of ten options is paralysis. Three is the maximum useful number; two is often better. - **Hide your recommendation behind neutrality.** "Either could work." If you have a view, state it. The operator wants pushback, not a menu. - **Confuse activity with progress.** Long lists of recent commits are not strategic information. Strategic information is "we did X, the signal moved Y, now Z is the constraint." - **Assume the strategy file is current.** Cross-grounding catches drift. The auto-generated state file may lag the actual state. ## Output shape Roughly: ``` ## Where we actually are ## The binding constraint ## Three lanes ### Lane A — ### Lane B — ### Lane C — ## My recommendation ## Question ``` Keep each section tight. The operator should be able to read the whole thing in 90 seconds. ## Calibration signals After the realignment, watch for: - The operator picks one of the named lanes — good; the framing was useful. - The operator picks a fourth lane you didn't surface — moderate; you missed an option but the conversation moved. - The operator pushes back on the framing itself ("you've got the constraint wrong") — useful; update and re-present. - The operator asks for more detail before deciding — fine; provide it scoped to the chosen lane. - The operator decides with no engagement — bad; the framing was probably wrong or trivial. Audit next time. ## Anti-patterns to refuse - **Realigning every week.** Realignment fatigue is real. If nothing material has changed, the right response to "what's next?" is "Lane X (from last realignment) is still the right answer; here's the next concrete move." - **Realigning to avoid the hard work.** Sometimes "what's next?" is dodging the current task. Notice if the operator's recent moves suggest they're stuck mid-execution; the right move may be helping them finish, not redirecting. - **Authoring the next arc immediately.** A realignment surfaces the shape; the actual arc plan is a separate document the operator commissions after the realignment confirms the direction. ## Pairs with - `decision-memo` — when the operator confirms a lane, the next move is often a decision memo on the specific change. - `memory-write` — capture the binding-constraint reframe so the next realignment starts from the new baseline, not the old one. - Any project's roadmap-amendment flow — if the realignment surfaces drift in the canonical strategy doc, draft an amendment. --- > Source: [tomevault-io/companyos](https://github.com/tomevault-io/companyos) — distributed by [TomeVault](https://tomevault.io).