Decision Evaluation Framework

Apply 20+ classical decision-making frameworks to any major decision — personal or professional — and produce a structured, multi-lens evaluation that you can archive, share, or upload to Google Drive as a polished PDF.

The plugin treats each framework as a distinct lens. Run a single lens, run a curated bundle of lenses (e.g. quick, strategic, personal, crisis), or run all 20 in parallel and let the synthesis aggregator surface convergence and divergence across them. The output is not a determinative answer — it's a structured, comparative view of the decision from multiple angles.

Frameworks shipped

Framework	Best for
`cost-benefit`	Decisions where outcomes are quantifiable
`pre-mortem`	Surfacing failure modes before commitment
`mcda`	Multi-option, multi-criterion comparison
`decision-tree`	Discrete outcomes with estimable probabilities
`reversibility`	Calibrating how much deliberation a decision deserves
`regret-minimization`	Long-horizon personal decisions
`ooda`	Fast-moving or adversarial situations
`eisenhower`	Prioritization decisions
`swot`	Strategic positioning
`second-order`	Decisions where consequences cascade
`opportunity-cost`	When the option set is too narrow
`ten-ten-ten`	Debiasing short-term emotion
`inversion`	Avoiding stupidity vs seeking brilliance
`base-rates`	Defeating inside-view optimism
`kepner-tregoe`	Procurement, vendor selection, formal alternatives analysis
`six-hats`	Group decisions, balanced deliberation
`cynefin`	Matching decision style to problem domain
`red-team`	Steel-manning the case against the leading option
`stakeholder-map`	Decisions with political/organisational dimensions
`time-horizon`	When short and long time-horizons may conflict

Skills

Workflow:

onboard — first-run setup. Writes config (output path, default frameworks, bundles, Drive settings).
new-decision — interview-driven scaffolding of a structured decision file.
analyze — orchestrator. Runs N frameworks in parallel as subagents.
synthesize — aggregates per-framework fragments into a cross-framework synthesis with score table and convergence/divergence analysis.
render-pdf — Typst-rendered PDF (cover, exec summary, per-framework sections, synthesis).
export — uploads PDF to Google Drive (via gws MCP) or copies to a local path.
bundles — manage named lens-bundles.

Per-framework lens skills (run individually or as building blocks): one skill per framework in the table above.

Quickstart

# Install
claude plugins install decision-evaluation-framework@danielrosehill

# In Claude Code:
/decision:onboard                          # one-time setup
/decision:new-decision                     # scaffold a decision file
/decision:analyze --bundle=strategic       # run the strategic bundle in parallel
/decision:synthesize <workspace>           # aggregate (auto-runs after analyze)
/decision:render-pdf <workspace>           # render to PDF
/decision:export <workspace> --drive       # upload to Drive

Or, for a single lens:

/decision:pre-mortem <decision.md>

Workspace layout

Each analysis lives in its own folder:

<output_root>/<YYYY-MM-DD>-<slug>/
  decision.md         # input decision
  fragments/          # one markdown per framework
    cost-benefit.md
    pre-mortem.md
    ...
  synthesis.md        # cross-framework synthesis
  analysis.pdf        # rendered output

Configuration

Stored at ${CLAUDE_USER_DATA:-${XDG_DATA_HOME:-~/.local/share}/claude-plugins}/decision-evaluation-framework/config.json. Edit directly or re-run /decision:onboard.

Philosophy

Frameworks are lenses, not oracles. Different frameworks are good at different things. A pre-mortem will catch failure modes that a cost-benefit analysis is structurally blind to; a regret-minimization frame will reach a different conclusion than an OODA frame on the same career decision, and that disagreement is the signal. The synthesis step doesn't average it away — it surfaces it so you can decide which lens deserves more weight given the specific decision in front of you.

License

MIT.