Persist agent eval metrics in Braintrust - #334

Merged

jlitola merged 27 commits into

mainfrom

jlitola/braintrust-eval-persistence

Sep 1, 2026

Commits on Aug 31, 2026

feat: retain eval prompt and answer evidence
jlitola
committed
2c81b12
feat: map eval suites to Braintrust rows
jlitola
committed
026dccf
feat: publish eval rows to Braintrust
jlitola
committed
cf0a956
feat: add Braintrust eval exporter CLI
jlitola
committed
a561f14
ci: persist agent eval metrics in Braintrust
jlitola
committed
f42582b
docs: document Braintrust agent eval operations
jlitola
committed
b8cd399
fix: clean Braintrust eval lint warnings
jlitola
committed
d0a655c
docs: correct Braintrust eval evidence
jlitola
committed
869be6e
fix: harden Braintrust eval preflight provenance
jlitola
committed
4668b48
docs: document empty suite preflight rejection
jlitola
committed
7fe935e
feat: map agent evals to native Braintrust metrics
jlitola
committed
93f58d2
docs: record native Braintrust readback
jlitola
committed
46861b1
feat: capture eval tool lifecycle timestamps
jlitola
committed
e9df4ce
feat: persist native Braintrust tool spans
jlitola
committed
4850299
fix: record Braintrust tool durations
jlitola
committed
2135d4b
docs: document native Braintrust tool spans
jlitola
committed
e1b9621
fix: reject inconsistent Braintrust tool telemetry
jlitola
committed
bfb4bdd
fix: tighten eval tool telemetry diagnostics
jlitola
committed
621a3b2
fix: reject retained terminal timing
jlitola
committed
7195ccc
docs: record labeled Braintrust eval proof
jlitola
committed
dd01bce
docs: plan Braintrust experiment linkage
jlitola
committed
47fe32c
feat: link Braintrust eval experiments
jlitola
committed
6e7d10a
ci: route Braintrust eval identity
jlitola
committed
578be55
docs: document Braintrust eval linkage
jlitola
committed
7539a89
fix: require Braintrust base readback
jlitola
committed
be5a037
test: clear Braintrust workflow lint
jlitola
committed
8a71a03
fix: report actual Braintrust experiment name
jlitola
committed
c773894

Navigation Menu

Uh oh!

Persist agent eval metrics in Braintrust - #334

Persist agent eval metrics in Braintrust#334
jlitola merged 27 commits into
mainfrom
jlitola/braintrust-eval-persistence

Commits on Aug 31, 2026

feat: retain eval prompt and answer evidence

feat: map eval suites to Braintrust rows

feat: publish eval rows to Braintrust

feat: add Braintrust eval exporter CLI

ci: persist agent eval metrics in Braintrust

docs: document Braintrust agent eval operations

fix: clean Braintrust eval lint warnings

docs: correct Braintrust eval evidence

fix: harden Braintrust eval preflight provenance

docs: document empty suite preflight rejection

feat: map agent evals to native Braintrust metrics

docs: record native Braintrust readback

feat: capture eval tool lifecycle timestamps

feat: persist native Braintrust tool spans

fix: record Braintrust tool durations

docs: document native Braintrust tool spans

fix: reject inconsistent Braintrust tool telemetry

fix: tighten eval tool telemetry diagnostics

fix: reject retained terminal timing

docs: record labeled Braintrust eval proof

docs: plan Braintrust experiment linkage

feat: link Braintrust eval experiments

ci: route Braintrust eval identity

docs: document Braintrust eval linkage

fix: require Braintrust base readback

test: clear Braintrust workflow lint

fix: report actual Braintrust experiment name