| name | plan-ceo-review |
| description | Review a product plan like a CEO/founder, challenge the premise, compare strategic approaches, and write a durable local report. |
| compatibility | opencode |
| metadata | {"host":"opencode","migration_phase":"7","mode":"report-first"} |
What This Skill Does
This is the first OpenCode-native /plan-ceo-review slice for gstack. It reviews an explicit product or strategy brief and writes a bounded CEO/founder-style planning report.
In this v1 slice:
- read a provided brief or brief file
- challenge whether the stated feature is the right problem to solve
- describe the 10x version that creates much more value
- compare at least two implementation approaches
- recommend what to build now versus later
- write a durable local report
Do not modify code, docs, or plans in this v1 workflow.
Required Input
You need either:
- an inline product brief, or
- a repo-local brief file like
PLAN.md
If no brief is provided, ask for exactly one thing: the brief text or the brief file path.
Execution Contract
- Stay local; do not load unrelated skills
- Do not use background tasks or todo lists for this v1 flow
- Write the report file before returning the final answer
- Keep the report bounded to the fixed sections below
- Do not require multi-turn interaction
- Do not call external services
- Do not modify repository files other than the requested report output
Output
Write the report under .gstack/plan-reports/.
Default path:
.gstack/plan-reports/plan-ceo-review-{slug}.md
Honor an explicit output path if the request provides one.
Fixed Report Sections
The report must include these sections:
Inputs Reviewed
Premise Challenge
10x Check
Alternative Approaches
Recommendation
Scope For This Slice
Deferrals
Local Validation
Not In Scope
Workflow
1. Read Inputs
- read the brief first
- read only lightweight local context files if they are named or obviously relevant
Good local context examples:
README.md
ARCHITECTURE.md
- the specific files named in the brief
2. Produce Premise Challenge
Challenge the brief directly and concretely.
Answer questions like:
- is this the right problem to solve?
- is the requested feature the real product, or just a proxy?
- what user or business outcome actually matters?
- what happens if we do nothing?
Keep this section concise and specific to the brief.
3. Produce 10x Check
Describe the version that creates dramatically more value without turning this into an unlimited-scope fantasy.
The 10x check should:
- stay anchored to the same user problem
- expand ambition, not vagueness
- make the better end state concrete
4. Produce Alternative Approaches
Include at least 2 distinct approaches every time:
- one minimal viable path
- one ideal long-term path
For each approach, include:
- summary
- effort
- risk
- pros
- cons
Prefer concrete tradeoffs over abstract architecture language.
5. Produce Recommendation
Choose one approach and explain why.
The recommendation must:
- state what to build now
- state what to defer
- explain why this is the right slice
- keep the first implementation credible for a small local OpenCode migration
6. Produce Scope and Deferrals
Scope For This Slice should name only the work that belongs in the immediate first migration slice.
Deferrals should capture the meaningful follow-up work that is intentionally excluded.
7. Produce Local Validation
Describe how to validate the slice locally.
Prefer local checks such as:
- temp repo smoke test
- explicit brief file like
PLAN.md
opencode run --command ...
- targeted Bun test commands
8. Produce Not In Scope
Explicitly list the important things this first CEO review slice does not do.
Rules
- Keep the output concrete, strategic, and product-centered
- Prefer explicit tradeoffs over inspirational language
- Do not invent product scope beyond what the brief justifies
- Do not require interactive multi-pass review in this v1 workflow
- Do not turn this into an engineering design review
- Do not modify repository files other than the requested report output