Package extracted findings into a compact, decision-ready format for engineers, managers, or review bodies. Use this after running complexity-mapper, architecture-risk-review, or context-sharding to synthesize findings into a structured brief with options, evidence, risks, and recommended next steps. Produces a Decision Brief with evidence summary, top risks, unresolved questions, and actionable recommendations. Ask for a "decision brief", "decision package", or "summarize findings for a decision" to use this workflow.
How this skill is triggered — by the user, by Claude, or both
Slash command
/systems-thinking-plugin:decision-briefThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Use Decision Brief when:
Use Decision Brief when:
Do not use Decision Brief when:
complexity-mapper, context-sharding, or architecture-risk-review first to generate the raw findings.| Input | Required | Description |
|---|---|---|
| Extracted findings | Yes | Outputs from prior workflow runs: Context Packets, Complexity Heat Maps, Hidden Risk Summaries, or direct analysis results. The more structured the input, the better the brief. |
| Ranked risks | Yes | Prioritized list of risks with severity assessments. Typically produced by complexity-mapper or architecture-risk-review. If not available, the brief will flag this gap. |
| Assumptions | Yes | Assumptions underlying the findings — both validated and unvalidated. These become critical context for decision-makers. |
| Decision audience | No | Who will read this brief and what decisions they are empowered to make. Shapes the level of technical detail and the framing of options. Defaults to technical leadership. |
| Options considered | No | If multiple design or implementation options were evaluated, provide them. If not provided, the brief will present findings as applying to the current proposed approach. |
Collect all available outputs from prior workflow runs or direct input:
context-shardingcomplexity-mapperarchitecture-risk-reviewpattern-remixweb-researcher (if web research was performed upstream)extraction-planner (for understanding how extraction was parallelized)For each input, verify:
If the inputs are insufficient for a meaningful brief — fewer than three substantive findings, no risk assessment, or no concrete evidence — stop and recommend running complexity-mapper or another appropriate upstream workflow first. Do not produce a thin brief that creates false confidence.
Before final synthesis, check reference/examples/ for sample Decision Brief outputs. Match the format and depth of any found examples to maintain consistency. Also check reference/prompts/ for decision-framing patterns the user has found effective — these can sharpen the brief's structure and language.
Pass all gathered inputs to the synthesis-brief-writer agent with the Decision Brief output contract and the following instructions:
Structure the brief to answer these questions in order:
Synthesis-brief-writer constraints:
[Inferred] so decision-makers can distinguish evidence from interpretation.Before presenting, check the Decision Brief against these requirements:
[Inferred].If any check fails, return to the synthesis-brief-writer with specific feedback.
Deliver the Decision Brief with:
The output is a Decision Brief conforming to the output contract:
## Decision Brief
### Decision Frame
[What is being decided, who is deciding, and what authority level is required]
### Options Considered
| Option | Description | Analysis Depth |
|---|---|---|
| [Option A] | [One-sentence description] | [Full / Partial / Surface-level] |
| [Option B] | [One-sentence description] | [Full / Partial / Surface-level] |
### Evidence Summary
#### Option A
- [Finding]: [Source] — Confidence: [H/M/L]
- [Finding]: [Source] — Confidence: [H/M/L]
- [Inferred] [Concern]: [Reasoning] — Confidence: [L]
#### Option B
- [Finding]: [Source] — Confidence: [H/M/L]
### Top Risks
| Risk | Severity | Trigger | Evidence Source | Confidence |
|---|---|---|---|---|
| [Risk] | [Critical/High/Medium] | [What causes it] | [Source anchor] | [H/M/L] |
### Contradictions (if any)
- [Source A] says [X]; [Source B] says [Y]. Resolution requires: [action].
### Unresolved Questions
| Question | Why It Matters | Suggested Resolution |
|---|---|---|
| [Question] | [Impact on decision] | [How to answer it] |
### Recommended Next Steps
1. [Action]: [Rationale] — Owner: [Suggested]
2. [Action]: [Rationale] — Owner: [Suggested]
### Appendices (if needed)
- [Appendix A: Full risk inventory]
- [Appendix B: Detailed evidence tables]
| Failure Mode | Signal | Response |
|---|---|---|
| Insufficient extraction data | Step 1 yields fewer than 3 substantive findings or no risk assessment | Stop. Do not produce a thin brief. Recommend running complexity-mapper or architecture-risk-review first to generate adequate raw material. |
| Single-perspective input | All findings come from one source, one extraction agent, or one analysis workflow | Flag in the brief that the evidence base is limited. Lower all confidence ratings by one level. Recommend additional analysis from a different angle. |
| Missing cost or dependency analysis | No findings related to cost implications or system dependencies | Flag the gap explicitly in the Unresolved Questions section. Recommend running the relevant extraction (cost-capacity-analyst or architecture-dependency-mapper) before finalizing the decision. |
| Stale findings | Input findings are based on documentation or analysis more than 30 days old | Flag recency concerns in the brief. Recommend re-running upstream analysis with current documentation before making the decision. |
| False precision | Brief presents low-confidence findings as high-confidence, or uses precise numbers from rough estimates | Review confidence ratings in Step 3. Ensure [Inferred] labels are applied. Remove false precision from quantitative claims. |
| Scope creep in synthesis | synthesis-brief-writer introduces new analysis not present in the inputs | Reject and re-invoke. The brief synthesizes; it does not analyze. New observations must be labeled [Inferred] at minimum, but the brief should not be the place for new investigation. |
npx claudepluginhub fakoli/systems-thinking-plugin --plugin systems-thinking-pluginProvides structured framework with specialized agents for research, analysis, architecture decisions, validation, synthesis, reviews, investigations, and reporting. For complex problems needing systematic exploration and persistent artifacts.
Produces concise decision-ready briefing notes from processed research materials, with structured sections for findings, sources, risks, and recommended actions.
Stress-tests design decisions with premortem/backcast analysis to produce risk-calibrated ADR recommendations. For multiple architecture options, irreversible changes, or complex trade-offs.