Strategic Brief
A pre-commitment ritual that forces demand validation before resources are committed.
Core insight: an adversarial review before commitment closes the gap between "this sounds like a good idea" and "this has real demand evidence". The default instinct is to execute on the first plausible idea. This skill forces the questions: what is missing, what is the strongest counterargument, and which assumption is most likely wrong?
Paths below are relative to this skill's directory; resolve them to absolute paths before reading anything.
When to use
- Before opening any tracked work item expected to take more than two days.
- Before green-lighting a new product feature, a strategic initiative, or a significant marketing investment.
- When a request arrives without clear demand evidence or success criteria.
- When the scope feels fuzzy, or the problem definition shifts from one conversation to the next.
When not to use
- Tactical execution with a clear specification: work the issue directly.
- Work whose scope the owner has already locked.
- Tasks under about two days, where the overhead is not warranted.
- A plan already drafted that needs a leadership read:
ceo-review. Several candidates to order:rice-prioritization. A campaign about to ship:campaign-brief.
Where the company context comes from
Demand evidence, the segment and "why now" are judged against what the company is betting on and who it sells to, so the brief needs that context and must not invent it.
This repository's agent configuration should have been provided to you (the
## Agent skills block pointing at docs/agents/). If it is missing, use the
defaults below and suggest the user run the setup-aclarify-skills skill.
- Context source. In order: a strategy, positioning, ideal-customer or
roadmap document the user names; one in the conventions folder named in
docs/agents/workspace.md(default: none); what the user pasted into the conversation. - With none of these, ask the user for the minimum: the two or three core bets the company is making now, and who it sells to. Ask the user, using your structured-question tool if you have one; otherwise list the questions numbered in plain text and stop until they reply. If they would rather you proceed, run on the request alone and mark Q1 and Q6 not checked against stated strategy.
- Issue tracker (
docs/agents/issue-tracker.md): where the brief and verdict are posted when the initiative lives in an issue. Default: none configured. - Scratch folder (
docs/agents/workspace.md): where the brief is saved when it has no other home. Default:.scratchpad/, git-ignored.
Step 1: the demand diagnostic
Ask the person who brought the idea these six questions in order, or reason through them yourself from the context you have. Do not move on until the current one has a specific answer. Generic answers ("our users", "improve engagement") are not acceptable: push for a named segment, concrete evidence and a measurable definition.
- Who specifically has this problem? Not "users" or "customers". Name the segment: a job title, a company type and a situation. For example: "head of revenue operations at a 50 to 150 person B2B software company, managing a team of three."
- What is the evidence of demand? Direct customer conversations (quote them), observed workarounds, support tickets, prior campaign data, a competitive gap analysis. "We think" and "it seems" are not evidence.
- What does success look like in 90 days? Measurable and specific. Not "improve retention", but "reduce 30-day churn from 8% to 5% in that segment."
- What is the minimal viable version? What would you cut with half the time? The answer names the true core bet and exposes scope creep in the original description.
- What are you explicitly not building? Force a scope boundary. If nothing is excluded, the scope is undefined.
- Why now? Why is this the right thing to work on today rather than the next best alternative? What does delay cost, and what does doing this instead of that alternative cost?
Step 2: write the initiative brief
Synthesize the answers into a brief:
# Initiative brief: [initiative name]
**Date:** [YYYY-MM-DD]
**Initiated by:** [person or role]
**Estimated scope:** [days of work]
## Problem
[Who has the problem, in specific terms. Two or three sentences.]
## Demand evidence
[Concrete evidence: quotes, data, observed behaviour. No speculation.]
## Success criteria (90 days)
[One measurable outcome, in one sentence.]
## Minimal viable scope
[What is being built, as a bulleted list.]
## Explicit non-scope
[What is not being built, as a bulleted list.]
## Why now
[The opportunity-cost reasoning, in two or three sentences.]
Step 3: adversarial review, up to two rounds
Challenge the brief on four dimensions using the prompt in references/adversary.md. If you can dispatch subagents, give that prompt and the brief to one fresh subagent: a reader who did not sit through the diagnostic has not already been persuaded by it. Otherwise work the prompt yourself, reading the brief as a sceptic who has not seen the conversation behind it.
Revise the brief against every challenge. If the revision answers them all, skip the second round. If challenges remain unresolved after two rounds, stop and put them to whoever owns the decision; do not proceed without resolution.
Step 4: go or no-go
## Go/no-go recommendation
**Verdict:** GO / NO-GO / CONDITIONAL GO
**Conditions (if any):**
- [A specific condition that must be resolved before proceeding]
**The one or two concerns that most need resolution:**
1. [The highest-risk open question]
2. [The second, if there is one]
- GO: the demand evidence is strong, the scope is clear, and the success criteria are measurable.
- NO-GO: the demand evidence is weak or missing. Do not proceed until it is gathered.
- CONDITIONAL GO: proceed only once the listed conditions are resolved.
Step 5: publish and link
- The initiative lives in an issue in the configured tracker: attach the
brief to it the way
docs/agents/issue-tracker.mddescribes, and post the verdict and conditions as a comment. - A new initiative is being opened: the brief goes in first, before any implementation sub-issues are created.
- No tracker configured: write the brief beside the planning document it belongs to, or where the user asks; if they want it kept out of tracked files, save it as a file in the scratch folder. Give the user the verdict and conditions to pass on to whoever needs them.
Key principle: specificity over generality
Named customers, specific examples, concrete data points, always. "Three customers in our target segment told us they spend four hours a week on this by hand" beats "customers find this valuable". The brief should be quotable to the people who fund the work.