Observed since 20 Oct, 21:57 UTC
No superseded metadata states collected. Unchanged observations do not create new revisions.
Evidently AI metrics + drift hook for AgentForge
Evidently AI agent metrics + drift hook for the AgentForge framework.
Implements the StepHook + FinishHook contracts and
registers as agentforge.hooks:evidently.
pip install agentforge-evidently[evidently]
from agentforge import Agent
from agentforge_evidently import EvidentlyHook
hook = EvidentlyHook.from_config(
project="my-agent",
report_dir="./evidently-reports",
)
agent = Agent(
model="bedrock:...",
on_step=hook,
on_finish=hook,
)
Per-step records (one row per Step):
| Column | Source |
|---|---|
run_id | bound RunContext.run_id |
iteration | step.iteration |
kind | step.kind |
cost_usd | step.cost_usd |
tokens_in / tokens_out | step.tokens_in / step.tokens_out |
duration_ms | step.duration_ms |
has_tool_call | step.tool_call is not None |
At finish, the hook appends a run-level row and writes an
Evidently Report JSON to
<report_dir>/<run_id>.json. The report carries the buffered
rows + RunResult.finish_reason + RunResult.cost_usd so an
offline analysis pipeline can compute drift / quality metrics
across runs.
Apache-2.0.
Download counts and popularity will appear here as the collector receives them.
| Version | Published | Direct dependencies | Open version |
|---|---|---|---|
| 0.4.0 | 24 Jun 2026 | 2 |
This describes our collected data; it is not a security assessment.
Observed since 20 Oct, 21:57 UTC
No superseded metadata states collected. Unchanged observations do not create new revisions.
package:e410b826f6f95a865f8eb1508fb504a391d222db00ce27150e8a18dfc3e46c39