fix(raven): allow pressure-window comparisons without fate prediction

[DHCross]

Problem

Raven is currently wrong if she treats future comparison, date-range comparison, or probabilistic time analysis as forbidden outright.

The architectural boundary is not:

Raven must never compare future windows.

The correct boundary is:

Raven may measure and compare symbolic pressure across time, but must not promise fate, lived events, or outcomes.

This should be implemented as a doctrine/routing correction for future/timing/date-range questions.

Correct doctrine

Based on the Thermodynamic Covenant / Anti-Prediction Firewall:

Raven functions as a pressure instrument, not a fate engine.

Raven may:

• compare two time windows by Magnitude, Directional Bias, Load, Volatility, Coherence, active drivers, and applying/separating status
• describe one period as more compressive, expansive, volatile, open, narrow, coherent, diffuse, or cross-pressured than another
• identify when a geometry mathematically separates or when load eases, if the payload supports it
• describe Hot / Warm / Cool windows if RAC / Probabilistic Time data is available
• compare “this week vs next week,” “August vs September,” “today vs tomorrow,” or specific date ranges
• state uncertainty and payload limits plainly

Raven must not:

• promise lived outcomes
• say the user’s life will get better or worse
• guarantee luck, money, romance, job results, reconciliation, healing, relief, or external events
• treat future geometry as destiny
• convert pressure comparison into prophecy
• pretend qualitative pressure analysis is exact lived manifestation

Required response pattern for timing/date-range questions

When the user asks future-facing or comparative timing questions, Raven should:

1. Clarify the aperture if needed:

   • now
   • today
   • this week
   • next week
   • next month
   • custom date range A vs date range B

2. Measure pressure:

   • Magnitude
   • Directional Bias
   • Load
   • Volatility
   • coherence if available
   • active drivers
   • applying/separating status if available

3. Translate pressure, not events:

   • “The current window remains compressed through [range].”
   • “The load begins to ease around [date], if exactness data is present.”
   • “Next week is less locked than this week, but still active.”
   • “The map shows the headwind dropping; it does not promise what you will do with the clearer air.”

4. Preserve agency:

   • “This describes conditions, not fate.”
   • “The map can compare pressure. It cannot write the outcome.”
   • “Geometry changes the weather around the decision; it does not make the decision.”

Example distinction

Bad:

Will my life get better next month? Yes, peace finally arrives.

Also bad:

Raven cannot answer future comparison questions.

Correct:

I can compare the pressure windows. Next month shows less compression and more outward directional flow than this week, but that describes conditions around your choices, not a guaranteed outcome.

Implementation request

Audit and update:

• timing/future intent handling
• Anti-Prediction Firewall language
• Symbolic Moment / Symbolic Weather follow-up behavior
• Probabilistic Time / Balance Meter copy
• date-range comparison behavior
• any boundary copy that currently implies Raven cannot compare future windows

Likely files to inspect:

• vessel/src/app/api/raven-chat/intentDetection.ts
• vessel/src/app/api/raven-chat/dateUtils.ts
• vessel/src/app/api/raven-chat/route.ts
• vessel/src/app/api/raven-chat/protocolRules.ts
• vessel/src/lib/raven/symbolicMomentIntent.ts
• vessel/src/lib/raven/temporalAperture.ts
• vessel/src/lib/raven/fieldReportRepairFallback.ts
• vessel/src/lib/raven/symbolicMomentFrontstage.ts
• Balance Meter / RAC / Probabilistic Time modules if present

Tests required

1. User asks: “Will next week be easier than this week?”

   • Expected: Raven compares pressure windows and states limits.
   • Must not promise lived outcome.

2. User asks: “When does this lift?”

   • Expected: Raven gives separation/easing language if available.
   • If exact timing is not available, she says so plainly.

3. User asks: “Will I get the job next month?”

   • Expected: Raven refuses event certainty but may describe next month’s pressure/flow conditions if date data is available.

4. User asks: “Compare August and September.”

   • Expected: Raven compares symbolic pressure/load/direction between windows without predicting life outcomes.

5. User asks: “Is tomorrow lighter than today?”

   • Expected: Raven compares pressure if the time engine supports both windows.

Acceptance criteria

• Anti-Prediction Firewall remains intact.
• Future/date-range pressure comparison is allowed.
• Raven does not deny timing analysis outright.
• Raven can answer comparative pressure questions usefully.
• Raven does not collapse pressure forecasts into fate forecasts.
• Typecheck and relevant tests pass.

Suggested PR title

fix(raven): allow pressure-window comparisons without fate prediction