unsure-not-sure-about-this

name	unsure - Not Sure About This
description	For mild uncertainty — you have a direction but aren't confident. A decision tree to move from hesitation to action or justified pause.
output	{"format":"prose"}

Unsure - Not Sure About This

Input: $ARGUMENTS

---

Core Principles

1. Uncertainty is information, not weakness. Being unsure means you're aware of risk, complexity, or missing data. That awareness is an asset. The goal isn't to eliminate uncertainty — it's to make it actionable.

2. Most uncertainty is resolvable. The question is whether it's worth resolving BEFORE acting or whether you can resolve it BY acting.

3. Confidence is not binary. You don't need to go from unsure to certain. You need to go from unsure to "sure enough to move."

4. The cost of waiting matters. Uncertainty sometimes justifies delay. But delay has its own cost — missed windows, accumulated anxiety, lost momentum. Weigh both sides.

---

Phase 1: Name the Uncertainty

[U1] WHAT_UNSURE_ABOUT: [state it precisely — not "I'm not sure" but "I'm not sure whether X because Y"]
[U2] CURRENT_LEAN: [which way are you leaning, even slightly?]
[U3] STAKES: [what happens if you're wrong?]

---

Phase 2: What Would Make You Sure?

[U4] CERTAINTY_SOURCE: What kind of input would resolve this?

Source	Example	Speed
Information	A fact you don't have yet	Minutes to days
Validation	Someone confirming your thinking	Hours to days
Experience	You've never done this before	Only gained by doing
Time	You need to see how things develop	Days to weeks
Gut check	You know but don't trust yourself	Already available

[U5] IDENTIFIED_SOURCE: [information | validation | experience | time | gut]

---

Phase 3: Decision Tree

Can you get that certainty quickly (< 1 day)?
├── YES → Go get it. Then decide.
│   [U6] ACTION: [specific step to get the information/validation/etc.]
│   → Return here after.
│
└── NO → Is the cost of being wrong recoverable?
    ├── YES (recoverable) → Proceed with a checkpoint.
    │   [U7] PROCEED_ACTION: [what to do now]
    │   [U8] CHECKPOINT: [when/how to check if this was right]
    │   [U9] REVERSAL_COST: [what it takes to undo if wrong]
    │   "Act, check, adjust. Don't wait for certainty you can't get cheaply."
    │
    └── NO (not easily recoverable) → Invest in certainty.
        [U10] INVESTMENT: [what it takes to become more sure]
        [U11] TIMELINE: [how long this investment takes]
        [U12] MINIMUM_CONFIDENCE: [what level of certainty justifies acting]
        "This is worth slowing down for. The cost of being wrong exceeds the cost of waiting."

---

Phase 4: Gut Check Override

Before following the tree, check:

[U13] GUT_SAYS: [what does your instinct say, separate from analysis?]
[U14] GUT_CONFLICT: [does your gut conflict with the decision tree result?]

If gut conflicts with the tree:

• Gut says GO but tree says WAIT → Check if you're avoiding the discomfort of uncertainty rather than responding to actual risk
• Gut says WAIT but tree says GO → Check if there's information your gut is processing that your analysis missed. Name it if you can.

---

Phase 5: Output

NOT SURE ABOUT THIS
===================

UNCERTAINTY: [precise statement of what's unsure]
LEANING: [current lean]
STAKES: [what's at risk]

WHAT WOULD HELP: [certainty source]
CAN GET IT QUICKLY: [yes/no]

RECOMMENDATION:
  [One of three paths:]
  a) GET THE INFO: [specific action] — then decide
  b) PROCEED WITH CHECKPOINT: [action] — check at [checkpoint]
  c) INVEST IN CERTAINTY: [what to do] — timeline [X]

GUT CHECK: [aligned / conflicting — with note if conflicting]

READY FOR:
- /dcp [decision] — if this is really a decision between options
- /aex [assumption] — if the uncertainty is about an assumption you're making
- /ht [hypothesis] — if you want to test your lean before committing

---

Failure Modes

Failure	Signal	Fix
Analysis paralysis	User keeps seeking certainty for a recoverable decision	Name the recoverability explicitly — "If wrong, you can [undo action] at [cost]"
False urgency	Pushing to act when stakes are genuinely high	Check stakes honestly — not everything is recoverable
Gut dismissal	Ignoring intuition because the tree says otherwise	Gut check is mandatory, not optional
Certainty theater	Going through motions of research without actually resolving uncertainty	Ask: "Will this information actually change your decision?"
Proxy uncertainty	Unsure about X but the real issue is Y	If the named uncertainty feels thin, dig one level deeper

---

Depth Scaling

Depth	Phases	Decision Tree	Gut Check
1x	Name + tree	Single path	Skip
2x	Name + tree + gut	Single path + gut	Include
4x	Full + multiple scenarios	All paths explored	Deep gut analysis
8x	Full + meta-analysis of uncertainty patterns	Full tree + second-order effects	Gut + pattern history

Default: 2x. These are floors.

---

Pre-Completion Checklist

• Uncertainty named precisely (not vague)
• Current lean identified (even if slight)
• Stakes assessed honestly
• Certainty source identified
• Decision tree followed to a recommendation
• Gut check performed and noted
• Recommendation is specific and actionable

---

Integration

• Use from: "I'm not sure", "I think so but...", "maybe I should", mild hesitation
• Routes to: /dcp (decision process), /aex (assumption examination), /ht (hypothesis testing)
• Differs from /idk: idk has no direction; unsure has a direction but lacks confidence
• Differs from /cnfsd: cnfsd doesn't understand something; unsure understands but doesn't trust
• Differs from /dcp: dcp compares defined options; unsure may not have clear options yet