threshold

THRESHOLD — Commitment Gateway

Not all decisions deserve the same depth. A CSS color choice and a database primary key choice are not the same decision — but without this skill, the model treats them with roughly equal cognitive investment.

This skill creates the gate between thinking and committing. It measures the reversal cost, enforces proportional exploration, and requires an exit plan before any high-cost commitment.

---

The Failure Mode You Must Recognize

You are about to give a quick, confident answer to a decision that:

• Cannot be easily undone (database schema, public API, data deletion)
• Affects many people or systems (blast radius extends beyond the immediate scope)
• Was not explored proportionally (one approach considered, zero alternatives evaluated)

Uniform depth — treating all decisions with the same level of analysis — systematically under-invests in the decisions that matter most.

---

The Protocol

Step 1 — CLASSIFY: Write the Reversal Cost

THRESHOLD CLASSIFICATION
────────────────────────────────────────
Decision:          [what is being decided — one sentence]

Reversal cost:     [Category 1 / 2 / 3 / 4]
  Category 1 — TRIVIAL (minutes, no coordination):
    Variable names, file organization, CSS values, config
  Category 2 — MODERATE (hours to days, some coordination):
    Indexes, internal API shapes, library versions, service boundaries
  Category 3 — EXPENSIVE (days to weeks, significant coordination):
    Database schemas, public API contracts, auth systems, data formats
  Category 4 — CATASTROPHIC (weeks to months, or impossible):
    Primary keys, data deletion, published standards, architecture at scale

Reversal timeline: [minutes / hours / days / weeks / months / never]
Blast radius:      [self / team / users / ecosystem]
────────────────────────────────────────

When uncertain between two categories, pick the higher one. The cost of overthinking a reversible decision is wasted time. The cost of underthinking an irreversible one is permanent damage.