ncreighton estimate-tokens
Use this skill to approximate the LLM token count of a text before an API call, for cost/budget guarding. Averages two estimators (~4 chars/token and ~0.75 tokens/word). Deterministic, offline, no tokenizer dependency. Not exact, but reliable for gating oversized prompts.
- Kind: analyzer | Niche: ai | Cost tier: free | Version: 0.1.0
Empire Integration
import sys
sys.path.insert(0, r'C:\D\Claude Code Projects\_SHARED')
import empire_bootstrap # wires all shared paths
Usage
As a CLI:
python run.py --text <string>
From a Python agent (ATEN runtime):
from skill_library import executor
result = executor.execute("estimate-tokens", {text=...})
Parameters
| Name | Type | Required | Description |
|---|---|---|---|
text |
string | yes | Text to estimate |
Returns
object — {"status": "ok"|"error", "result": ...}