Documentation Index
Fetch the complete documentation index at: https://docs.canthus.org/llms.txt
Use this file to discover all available pages before exploring further.
Worked examples
Example A - Sarah neutral baseline
Example A - Sarah neutral baseline
Profile: Coefficient 2.6, typical pool ~35 manaCheck-in:
Typical day total is ~31 mana, leaving a small margin. Light cooking exceeds pool and is not surfaced.
[3, 3, 3] (axis ~1.0)| Task | relativeCost | Duration | Axis | Cost |
|---|---|---|---|---|
| Taking medication | 0.3 | 1 min | 1.0 | 0.4 |
| Brushing teeth | 1.0 | 2 min | 1.0 | 2.6 |
| Getting dressed | 2.3 | 5 min | 1.0 | 15 |
| Shower | 1.0 | 10 min | 1.0 | 13 |
| Light cooking | 1.5 | 20 min | 1.0 | 39 |
Example B - Sarah rough-day inflation
Example B - Sarah rough-day inflation
Profile: Coefficient 2.6, typical pool ~35 manaCheck-in:
Essentials remain visible even if over pool.
[2, 2, 2] (axis factor ~1.25)| Task | Neutral | Rough day | Change |
|---|---|---|---|
| Brushing teeth | 2.6 | 3.3 | +25% |
| Getting dressed | 15 | 19 | +25% |
| Shower | 13 | 16 | +25% |
Example C - James severe profile capacity
Example C - James severe profile capacity
Profile: Coefficient 4.0, typical pool ~17 mana
Meds + sitting up + teeth = 13.6 mana. Phone call generally does not fit unless something else is skipped.
| Task | relativeCost | Duration | Cost |
|---|---|---|---|
| Taking medication | 0.3 | 1 min | 0.6 |
| Sitting up in bed | 0.3 | 15 min | 9 |
| Brushing teeth | 1.0 | 2 min | 4 |
| Brief phone call | 0.8 | 5 min | 8 |
Example D - Post-recalibration (before)
Example D - Post-recalibration (before)
Snapshot: Coefficient remains
2.6 before recalibration.| Task | Formula | Result |
|---|---|---|
| Shower | 1.0 x 10 x 2.6 x 1.0 x 0.5 | 13 mana |
Example E - Post-recalibration (after)
Example E - Post-recalibration (after)
Coefficient improves from 2.6 to 2.0. Stored
relativeCost remains valid; no task backfill required.| Case | Formula | Result |
|---|---|---|
| Shower before | 1.0 x 10 x 2.6 x 1.0 x 0.5 | 13 mana |
| Shower after | 1.0 x 10 x 2.0 x 1.0 x 0.5 | 10 mana |
Example F - Return after absence confidence moderation
Example F - Return after absence confidence moderation
Pre-absence: coefficient 2.6, confidence 0.8, pool ~35.Day 1 back (rough check-in):
- Confidence modifier:
0.8 x 0.85 = 0.68 - Bias:
0.80 + 0.68 x 0.17 = 0.916 - Pool shown:
35 x 0.916 ~= 32
- Confidence modifier:
0.8 x 0.95 = 0.76
- Confidence restored to pre-absence level.
Example G - Deadline outside horizon stays gated
Example G - Deadline outside horizon stays gated
Setup: Deadline task due on Mar 20 from a Mar 7 compute day, with over-pool cost.Expected behavior: Task remains hidden when outside the computed horizon window and over remaining pool.
Example H - Deadline inside horizon is surfaced
Example H - Deadline inside horizon is surfaced
Setup: Deadline task due on Mar 8 from a Mar 7 compute day, with over-pool cost.Expected behavior: Task is surfaced because deadline horizon grants pool exemption.
Example I - Scheduled reserve warning
Example I - Scheduled reserve warning
Setup: Scheduled task lands within 7-day lookahead and preview cost exceeds
typicalPool x 0.20.Expected behavior: scheduledReserve notice appears with per-task preview mana and total preview mana.Example J - Missed temporal items remain actionable
Example J - Missed temporal items remain actionable
Setup: One missed deadline task and one missed scheduled task.Expected behavior: Both remain surfaced with neutral, non-shaming language and actionable completion path.
Example K - Seeded irregular timeline (moderate)
Example K - Seeded irregular timeline (moderate)
Setup: Deterministic synthetic moderate profile with mixed check-ins and completion history.Expected behavior: Stable deterministic output for pricing, pool, and surfaced list under irregular history.
Example L - Seeded irregular timeline (severe)
Example L - Seeded irregular timeline (severe)
Setup: Deterministic synthetic severe profile with mixed check-ins and completion history.Expected behavior: Deterministic low-capacity behavior, including tighter surfacing and conservative pool outputs.
Deterministic Harness Mapping
CI runs a deterministic harness and publishes JSON output atapp/build/mana_scenarios/report.json.
| Scenario ID | Source | Seed |
|---|---|---|
example-a-neutral | Example A neutral baseline | 1101 |
example-a-rough | Example B rough-day inflation | 1102 |
example-b-severe | Example C severe profile capacity | 1201 |
example-c-post-recalibration-before | Example D recalibration before | 1401 |
example-c-post-recalibration-after | Example E recalibration after | 1402 |
example-d-return-after-absence | Example F return-after-absence | 1301 |
example-e-deadline-outside-horizon | Example G deadline outside horizon | 1501 |
example-f-deadline-inside-horizon | Example H deadline inside horizon | 1502 |
example-g-scheduled-reserve-warning | Example I scheduled reserve warning | 1503 |
example-h-missed-temporal-neutral | Example J missed temporal neutral handling | 1504 |
seeded-irregular-mixed-1 | Example K seeded irregular moderate | 9001 |
seeded-irregular-mixed-2 | Example L seeded irregular severe | 9002 |
app/tool/mana_scenario_harness.dart