Repository navigation

feat(evals): mock the dart MCP server and make green-gate a reliable regression check - #168

Merged

ryzizub merged 15 commits into

mainfrom

session/stoic-pelican-3u8q

Oct 5, 2026

Commits on Oct 5, 2026

feat(evals): add the dart MCP mock

ryzizub
and
claude
committed
e25b0b0
feat(evals): seed the fixture with source and make the test mock green

ryzizub
and
claude
committed
f43052e
feat(evals): add a green-gate case that drives the four gates

ryzizub
and
claude
committed
1c16f3e
docs(evals): document the dart mock and fix the counts it made stale

ryzizub
and
claude
committed
328ad6c
fix(evals): address review findings on the dart mock change

ryzizub
and
claude
committed
f978501
docs(evals): correct three NOTES that still say no MCP server is available

ryzizub
and
claude
committed
0cfd68c
fix(evals): address review findings on the green-gate case and CI step

ryzizub
and
claude
committed
ccb5260
revert(ci): drop the mock manifest/body consistency check

ryzizub
and
claude
committed
e70e5aa
docs(evals): drop the volatile counts from the NOTES this branch touched

ryzizub
and
claude
committed
a6420eb
fix(evals): make the mock set agree with the fixture, and fix the rubric

ryzizub
and
claude
committed
8c26f49
fix(hooks): stop the session notice from asserting the MCP server is down

ryzizub
and
claude
committed
b1d4fc2
fix(evals): repair four rubrics that failed better-than-required answers

ryzizub
and
claude
committed
dbfa890
fix(evals): grade green-gate mechanically where the check is mechanical

ryzizub
and
claude
committed
3862a84
docs(evals): remove the per-skill NOTES.md files

ryzizub
and
claude
committed
0ff08ef
docs(evals): count is 102 after rebasing onto #165

ryzizub
and
claude
committed
3597224

Navigation Menu

Repository navigation

feat(evals): mock the dart MCP server and make green-gate a reliable regression check - #168

feat(evals): mock the dart MCP server and make green-gate a reliable regression check#168
ryzizub merged 15 commits into
mainfrom
session/stoic-pelican-3u8q

Commits on Oct 5, 2026

feat(evals): add the dart MCP mock

feat(evals): seed the fixture with source and make the test mock green

feat(evals): add a green-gate case that drives the four gates

docs(evals): document the dart mock and fix the counts it made stale

fix(evals): address review findings on the dart mock change

docs(evals): correct three NOTES that still say no MCP server is available

fix(evals): address review findings on the green-gate case and CI step

revert(ci): drop the mock manifest/body consistency check

docs(evals): drop the volatile counts from the NOTES this branch touched

fix(evals): make the mock set agree with the fixture, and fix the rubric

fix(hooks): stop the session notice from asserting the MCP server is down

fix(evals): repair four rubrics that failed better-than-required answers

fix(evals): grade green-gate mechanically where the check is mechanical

docs(evals): remove the per-skill NOTES.md files

docs(evals): count is 102 after rebasing onto #165