Launch energy recovery data agents to populate selfEnergy data files. Use when the user asks to generate, update, or review self-energy recovery data for the ER calculator.
Launch specialized agents to populate src/data/ercalc/selfEnergy/<region>.json with per-character energy recovery effects. Each agent reads character kits via impl_audit.py and writes to its own region file.
Note: This skill is specifically for the energy recovery data pipeline. For other ER calculator work, use standard subagents.
Each region has its own file under src/data/ercalc/selfEnergy/:
| File | Region |
|---|---|
mondstadt.json | Mondstadt |
liyue.json | Liyue |
inazuma.json | Inazuma |
sumeru.json |
| Sumeru |
fontaine.json | Fontaine |
natlan.json | Natlan |
nod-krai.json | Nod-Krai |
snezhnaya.json | Snezhnaya |
none.json | Travelers |
Before launching agents, run the candidate finder:
uv run --project scripts/pyproject.toml scripts/gen_self_energy.py
Reads each candidate's full kit, determines energy recovery mechanics, and writes correct JSON entries to the region file.
Launch (region scope):
Read `.claude/agents/energy-review.md` and follow its instructions. Region: mondstadt
Launch (entity scope):
Read `.claude/agents/energy-review.md` and follow its instructions. Entities: raiden_shogun, dori
Typical full sweep: 7-8 agents in parallel.
All region agents are fully independent — each writes to its own file. Launch all in parallel.
gen_self_energy.py to get candidates