Core game mechanic reference for the expedition momentum system. The Gradient Gauge tracks consecutive success/failure and scales issue difficulty accordingly. Applicable when the user asks about "gradient gauge", "expedition difficulty", "momentum system", "gradient level", or needs to understand how issue priority selection adapts based on past performance.
The Gradient Gauge is a momentum-based mechanic that adapts expedition difficulty based on recent performance. Inspired by the combat system of Clair Obscur: Expedition 33.
Stored in gradient.json (git-untracked, same directory as journal.tsv):
{
"level": 0,
"consecutive_failures": 0,
"consecutive_skips": 0,
"last_status": "",
"last_updated": ""
}
| Level | Name | Issue Priority Hint |
|---|---|---|
| 0 | Depleted | Only pick low-priority, small-scope issues |
| 1 | Warming | Low to normal priority |
| 2 | Steady |
| Normal priority |
| 3 | Rising | Normal to high priority |
| 4 | Surging | High priority OK |
| 5 | Gradient Attack | Attempt the most complex/highest-priority issue available |
level = min(level + 1, 5)consecutive_failures = 0consecutive_skips = 0level = max(level - 1, 0)consecutive_failures = 0consecutive_skips += 1level = 0consecutive_failures += 1consecutive_skips = 0When consecutive_failures >= 3 OR consecutive_skips >= 3:
gommage.md insight file with:
When querying Linear issues, filter by priority based on current gradient level:
Linear priority values: 0=No priority, 1=Urgent, 2=High, 3=Medium, 4=Low
Include the current gradient state in the expedition agent prompt:
Gradient Gauge: Level <N>/5 (<name>)
Priority hint: <hint text from table above>
Consecutive failures: <N> | Consecutive skips: <N>
Read and update gradient state:
# Read current state
python3 ${CLAUDE_PLUGIN_ROOT}/scripts/gradient.py read gradient.json
# Update after expedition result
python3 ${CLAUDE_PLUGIN_ROOT}/scripts/gradient.py update gradient.json <status>
Output is JSON for programmatic use.