| name | diagnose-strategy |
| description | Critique an existing strategy document using Richard Rumelt's good-strategy/bad-strategy diagnostic. Checks for the kernel (diagnosis + guiding policy + coherent action) and flags the four hallmarks of bad strategy: fluff, failure to face the challenge, mistaking goals for strategy, and bad strategic objectives. Use to pressure-test a product or company strategy before it ships. Read-only critique skill, not an authoring skill. |
| argument-hint | [path to the strategy document to diagnose] |
| user-invocable | true |
| allowed-tools | Read, Glob, Grep |
Diagnose a strategy (Rumelt good-vs-bad)
Critique the strategy document at $ARGUMENTS using Richard Rumelt's framework from Good Strategy / Bad Strategy (2011). Most documents called "strategy" aren't — they're goals, aspirations, or buzzwords arranged to look like a plan. This skill checks for the kernel of good strategy and flags the hallmarks of bad strategy, with evidence cited from the document.
This is a critique skill, not an authoring skill. It reads an existing strategy and reports findings. It does not rewrite. Strategy authoring belongs to /cpo:write-product-strategy; this skill tells you whether what was authored is actually a strategy. Run it as the mandatory follow-up after writing a strategy, or over any inherited strategy document.
Step 1: Read the document and identify what it claims to be
Read the full document. Establish:
- What it calls itself — strategy, plan, roadmap, vision, OKRs. The label sets the expectation you're testing against.
- The intended scope and horizon — product, company, portfolio; this quarter, this year, multi-year.
- The structure — what sections exist. Note immediately whether there is anything resembling a diagnosis, a guiding policy, and coherent actions; their presence or absence is the core of the assessment.
Output of this step: a one-paragraph summary of what the document is and how it's organised. Quote the document's own framing.
Step 2: Test for the kernel — diagnosis
Good strategy has a kernel of three interconnected elements. The first is the diagnosis: it simplifies a complex situation by naming the critical aspects and the biggest obstacle. Without a diagnosis, the rest of the strategy has nothing to address.
Check:
- Does the document name the central challenge or obstacle plainly? Or does it jump straight to goals and actions?
- Is the diagnosis a simplification that brings clarity, or a list of everything wrong (which is not a diagnosis)?
- Does it identify the critical few aspects, or treat all problems as equal?
Verdict: Present and sharp / Present but vague / Missing. Quote the diagnosis (or note its absence with the line where it should be).
Step 3: Test for the kernel — guiding policy
The guiding policy is the overall approach for overcoming the obstacle named in the diagnosis. It constrains and directs without specifying every action — like the strategy of a military campaign versus the individual orders.
Check:
- Is there an overall approach that follows from the diagnosis? ("Solve for the enterprise buyer, not the end user" is a guiding policy; "improve everything" is not.)
- Does it rule things out? A guiding policy that permits every action isn't guiding anything.
- Does it connect to the diagnosis? A guiding policy addressing a different problem than the one diagnosed is incoherent.
Verdict: Present and directive / Present but permits everything / Missing. Quote it or note the gap.
Step 4: Test for the kernel — coherent action
Coherent actions are the concrete, coordinated steps that carry out the guiding policy. The test is coherence: the actions reinforce one another rather than pulling in different directions or sitting as an unconnected list.
Check:
- Are there specific, coordinated actions, or just goals and targets?
- Do the actions reinforce each other, or compete for the same resources and attention?
- Do they implement the guiding policy, or are they a wish list assembled independently of it?
- Is there a proximate objective — a near-term target that is both achievable and real progress — rather than only distant aspirations?
Verdict: Coherent / Present but uncoordinated / Missing. Cite examples.
Step 5: Flag the four hallmarks of bad strategy
Independently of the kernel, scan for Rumelt's four signatures of bad strategy. Each found instance is a flag with a quote.
- Fluff — inflated language, abstraction, and buzzwords masking absence of content. ("Leverage synergies across our ecosystem to deliver world-class value.") Quote every instance. Dense fluff is the strongest single signal that the document is avoiding the actual challenge.
- Failure to face the challenge — no clear definition of the obstacle, so the strategy can't be evaluated against it. If Step 2 found no diagnosis, this hallmark is present by definition.
- Mistaking goals for strategy — statements of desire ("grow to $1B revenue", "become the category leader", "double our user base") presented as the plan. Goals describe the destination; strategy is how you get there. List each goal-as-strategy.
- Bad strategic objectives — either a scattered "dog's dinner" of unrelated goals, or "blue-sky" objectives disconnected from current capabilities. Note whether the objectives are a coherent set or a grab-bag, and whether they're reachable from where the organisation actually is.
Step 6: Score and produce the verdict
Synthesise into an overall verdict. The kernel is necessary: a document missing any of diagnosis, guiding policy, or coherent action is not a complete strategy, regardless of how polished it reads.
Overall rating:
| Rating | Meaning |
|---|
| Good strategy | Kernel complete and coherent; no significant bad-strategy hallmarks |
| Incomplete | Kernel has gaps (one element missing or vague); fixable with targeted work |
| Bad strategy | Multiple hallmarks present; kernel largely absent; mostly goals and fluff |
For each gap and flag, give a specific, actionable recommendation — what to add or fix, and which authoring skill produces it (/cpo:write-product-strategy for the strategy itself, /cpo:write-product-vision if the missing piece is actually the vision). Don't rewrite the strategy here; point to where the fix happens.
Rules
- Always cite the document. Every verdict and flag quotes the relevant line, or names the line/section where the missing element should be. "No diagnosis" is not a finding; "No diagnosis — the document opens at line 12 with goals; no obstacle is named anywhere" is a finding.
- Never rewrite the strategy. This skill critiques. It is read-only. Fixes happen in the authoring skills. Recommend; don't author.
- The kernel is necessary, not optional. A beautifully written document with no guiding policy is still incomplete. Don't let polish substitute for the kernel.
- Fluff is a signal, not a style note. Dense buzzword language usually means the author is avoiding a challenge they can't face. Flag it as evidence of that avoidance, not just as wording.
- Distinguish goals from strategy every time. "Grow to $X", "be the leader", "delight customers" are goals. If the document's "strategy" section is a list of these, say so plainly.
- Don't grade on a curve. A document can be the best strategy in the organisation and still be bad strategy. Rate against Rumelt's bar, not against what's normal locally.
Output Format
Return the critique directly (read-only skill — do not write a file unless asked).
# Strategy diagnosis — [document name]
**Document:** [path] · **Claims to be:** [strategy/plan/etc.] · **Scope:** [product/company] · **Date diagnosed:** [YYYY-MM-DD]
## Overall verdict: [Good strategy | Incomplete | Bad strategy]
[Two or three sentences summarising the assessment.]
## The kernel
| Element | Verdict | Evidence |
|---|---|---|
| Diagnosis | [Present and sharp / Present but vague / Missing] | [quote or line reference] |
| Guiding policy | [Present and directive / Permits everything / Missing] | [quote or line reference] |
| Coherent action | [Coherent / Uncoordinated / Missing] | [quote or line reference] |
## Bad-strategy hallmarks
| Hallmark | Found? | Evidence |
|---|---|---|
| Fluff | [Yes/No] | [quoted instances or "none found"] |
| Failure to face the challenge | [Yes/No] | [evidence] |
| Mistaking goals for strategy | [Yes/No] | [each goal-as-strategy quoted] |
| Bad strategic objectives | [Yes/No] | [scattered or disconnected objectives] |
## Recommended actions
1. [Highest-priority gap → what to add → which skill produces it]
2. [Next]
3. [Next]
Related skills
/cpo:write-product-strategy — authors the strategy this skill critiques; the place to fix any kernel gaps found here./cpo:write-product-vision — if the diagnosis reveals the missing element is actually the vision, write it here first.
