Make Decision

Goal

Help users make better decisions in minutes instead of hours by applying proven frameworks, detecting cognitive biases, and structuring the evaluation process.

make-decision

Comprehensive decision-making framework for structured evaluation of options. Contains 10 decision frameworks, 8 decision type classifications, 12 cognitive biases with debiasing strategies, 10 analysis techniques, 15 criteria templates (technology, tech stack, hiring, vendors, investment, market entry, product features, pricing, organizational change, location, job offers, relocation, education, housing, and a general fallback) and 8 group facilitation techniques. The plan names the options it found in the request, picks the framework, analysis methods and biases for the decision type, and suggests five weighted criteria. A scoring calculator totals the matrix, names the winner and finds the smallest weight change that would flip it.

Prerequisites

IMPORTANT: Detect the correct Python command first. Some systems use python3, others use python. Run:

python3 --version 2>/dev/null || python --version

Use whichever command succeeds (python3 or python) for ALL script calls below. If the system only has python (common on Windows), substitute python everywhere you see python3 in this document. The scripts need Python 3.9+ and only the standard library.

If neither works, Python is not installed: tell the user that the scripts need Python 3.9+ and ask before installing anything; do not run a package manager on your own.

Note: On Windows, Python 3 is typically available as python (not python3).

Running the Scripts

Run every command from the project root with the path shown, e.g. python3 .agents/skills/make-decision/scripts/search.py ... (the installer adjusts this path for your AI tool). Saved plans, workspaces and journals are written to the project (or to -o <dir>), never inside the skill folder. Saving again never overwrites files that already exist (they hold your notes); add --force to replace them.

Passing the user's text

Anything taken from the user's message goes on stdin with --stdin, never on the command line: the decision for --plan, the options for --matrix, the statement for --journal and the outcome for --journal --update. Quotes, backticks and $ then never reach the shell. Only short values you compose yourself (-c criteria, --scores, -p names) go on the command line.

Use exactly this delimiter, THINK_BETTER_EOF_7f3a, quoted as shown:

python3 .agents/skills/make-decision/scripts/search.py --stdin --plan <<'THINK_BETTER_EOF_7f3a'
<the user's text, unchanged>
THINK_BETTER_EOF_7f3a

Check the text first. The heredoc ends at the first line that is exactly THINK_BETTER_EOF_7f3a; anything after it would run as shell commands. If a line of the user's text is exactly that delimiter, do not use the heredoc: write the text unchanged to a temporary file with your file-editing tool (not the shell), run python3 .agents/skills/make-decision/scripts/search.py --stdin --plan < <file>, then delete the file.

In PowerShell (keep '@ at the start of its line):

$OutputEncoding = [Text.UTF8Encoding]::new()
@'
<the user's text, unchanged>
'@ | python .agents/skills/make-decision/scripts/search.py --stdin --plan

The here-string ends at a line that starts with '@. If a line of the user's text starts with '@, write the text to a file instead and run Get-Content -Raw -Encoding UTF8 <file> | python .agents/skills/make-decision/scripts/search.py --stdin --plan.

How to Use This Workflow

When user requests decision-making help (decide, choose between options, weigh a trade-off), follow this workflow. To find why something went wrong use problem-solving-pro; for bugs and code changes use code-solving.

Language: Respond in English. The scripts' output is in English: show it as it is, and keep commands, flags, file names and option names exactly as written (option names as the user wrote them).

If the user has not said what they are deciding, ask what the decision and the options are before running anything.

Step 1: Understand the Decision

Extract key information from user's decision description:

• Decision type: Binary choice, multi-option, resource allocation, strategic, operational, under uncertainty, group/stakeholder, time-pressured
• Options: What alternatives are being considered? (Add "do nothing" / "wait" if it is a real option.)
• Context: Industry, stakes, timeline, stakeholders, reversibility
• Constraints: Budget, time, resources, dependencies

Step 2: Generate Decision Plan (REQUIRED)

Always start with --plan to get comprehensive recommendations:

python3 .agents/skills/make-decision/scripts/search.py --stdin --plan -f markdown [-p "Project Name"] <<'THINK_BETTER_EOF_7f3a'
<the user's decision, unchanged>
THINK_BETTER_EOF_7f3a

If the user asked to save the work ("save", "step-by-step", "workspace"), run the Step 2b command instead of this one: it prints the same plan.

Classify it yourself when you can — you understand the decision better than keyword matching. Add --type "<decision type>", one of: Binary Choice, Multi-Option Selection, Resource Allocation, Strategic Direction, Operational / Tactical, Decision Under Uncertainty, Group / Stakeholder Decision, Time-Pressured Decision. Without it the script scores each type's signal phrases (e.g. "deadline", "board", "uncertain") and adds the options it found ("A vs B" or "A or B" leans Binary Choice, three or more lean Multi-Option Selection). The plan says what matched; when nothing did, it says so and lists the --type values: then re-run with --type.

The plan contains:

1. The request, the decision type (and why), and the options found in the request
2. The framework the decision type recommends, with its steps
3. Five weighted criteria from the best matching template (General Decision when none fits)
4. Analysis techniques and bias warnings chosen for the decision type and the domain
5. Facilitation for group decisions, anti-patterns, and a checklist
6. A Next Steps table at the end: show it once, do not add your own

Depth changes what the plan contains (--depth, default standard):

--json prints the plan as JSON (with options, criteria.items, bias_warnings, reversibility, ...).

Step 2b: Save the Plan, Resume Later

When the user asks to save, run this instead of the Step 2 command, not after it:

python3 .agents/skills/make-decision/scripts/search.py --stdin --plan --persist --step-docs -p "Project Name" -f markdown <<'THINK_BETTER_EOF_7f3a'
<the user's decision, unchanged>
THINK_BETTER_EOF_7f3a

--persist alone writes decision-plans/<project-name>/PLAN.md. With --step-docs it creates a workspace: 00-OVERVIEW.md (the request, options, and a step table with a Done? column), 01-DECISION-TYPE.md (the request quoted, decision statement, owner, deadline), 02-FRAMEWORK.md, 03-CRITERIA.md (the five criteria and weights pre-filled), 04-ANALYSIS.md, 05-OPTIONS.md (the options and a criteria × options matrix pre-filled), 06-DECISION.md (decision, pre-mortem, kill criteria, review date), BIAS-WARNINGS.md and DECISION-LOG.md.

In a later session (/decide.resume):

# Which steps are done and what to do next (-p name, or the workspace the text matches, or the latest)
python3 .agents/skills/make-decision/scripts/search.py --stdin --status [-p project-name] <<'THINK_BETTER_EOF_7f3a'
<the user's text, or nothing>
THINK_BETTER_EOF_7f3a
# Mark a step done (1-6 or classify, framework, criteria, analysis, options, decide); --undone reopens it
python3 .agents/skills/make-decision/scripts/search.py --done criteria -p project-name

Only mark a step done when the user has actually settled it.

Step 3: Deep-Dive Domain Searches

Use when the plan's recommendation needs more detail, OR when user asks about a specific topic (e.g., "what biases should I watch for?"). Pass the keywords on stdin like any other text; fixed example keywords such as those below may go on the command line:

python3 .agents/skills/make-decision/scripts/search.py --stdin --domain <domain> [-n <max_results>] <<'THINK_BETTER_EOF_7f3a'
<keywords>
THINK_BETTER_EOF_7f3a

Step 4: Compare and Score Options

With the options from the user's message on stdin:

# Empty matrix with the template's criteria and weights
python3 .agents/skills/make-decision/scripts/search.py --stdin --matrix -f markdown <<'THINK_BETTER_EOF_7f3a'
Which CRM: Salesforce, HubSpot or Pipedrive
THINK_BETTER_EOF_7f3a

# Your own criteria and weights, and the scores (1-5, one per criterion, in -c order)
python3 .agents/skills/make-decision/scripts/search.py --stdin --matrix -f markdown \
  -c "Cost:3,Speed:2,Risk:1" --scores "React:4,3,5;Vue:5,4,3" <<'THINK_BETTER_EOF_7f3a'
React vs Vue
THINK_BETTER_EOF_7f3a

Options are read from "A vs B vs C", "A, B or C", "Which X: A, B or C", "between A and B", "choose between A, B and C", "compare A with B". With --scores the matrix shows the weighted totals, the winner, and the sensitivity: for each criterion, the weight at which another option would tie the winner, with the smallest such change called out. A winner that flips under a small change is fragile: firm up that criterion before deciding. Weights may be any positive numbers (shown with their share of the total); -c "Cost,Speed" weighs them equally.