| name | decide |
| description | End-to-end decision support orchestrator. Compares options, runs risk
analysis on the recommended path, runs a bias check on the user's stated
reasoning, and produces a single decision brief. Use when the user says
"help me decide", "I need to make a call on this", "walk me through this
decision", "I'm stuck between these", or shares a real decision they're
about to commit to. Calls `comparative_analysis`, `analyze_risk`, and
`identify_cognitive_biases` in sequence.
|
| metadata | {"author":"Troy Taylor","version":"1.0","pattern":"orchestrator"} |
Decide
What This Skill Does
Runs the full Decision Duck pipeline for a real decision: compare options →
risk-check the recommended path → bias-check the user's reasoning →
assemble a one-page decision brief. Use this when the user is about to
commit, not when they're brainstorming.
When to Activate
- "Help me decide between …"
- "I need to make a call on …"
- "Walk me through this decision"
- "I'm leaning toward X — what should I check before committing"
- User shares 2+ options AND mentions an upcoming deadline or decision point
When NOT to Activate
- Just one tool's worth of value is needed → use the focused skill
- User wants to imagine the project failing → use
pre-mortem
- User wants an adversarial critique → use
red-team
Workflow
-
Gather the four inputs. Decision, options, criteria, current lean and why.
-
Step 1 — Compare.
comparative_analysis(
context: "<decision in one paragraph>",
options: [...],
criteria: [...]
)
-
Step 2 — Risk the recommended path.
analyze_risk(
scenario: "Committing to <recommended option> for <decision>",
context: "<one-paragraph decision background>"
)
-
Step 3 — Bias-check the user's reasoning.
identify_cognitive_biases(
analysis: "<the user's current lean and why, in their own words>"
)
-
Step 4 — Assemble the decision brief. One page. Lead with the
recommendation; the user should be able to skim the top half and decide.
-
Close with one decision-quality check. Pin the question that
sharpens the call.