Repository navigation
Conversation
🧙 Wizard CIRun the Wizard CI and test your changes against wizard-workbench example apps by replying with a GitHub comment using one of the following commands: Test all apps:
Test all apps in a directory:
Test an individual app:
Show more apps
Test against a Context Mill branch:
Add Results will be posted here when complete. |
1a9a29c to
a162b69
Compare
616023a to
50c996d
Compare
a162b69 to
c92c697
Compare
50c996d to
5fccf03
Compare
5fccf03 to
a777dfd
Compare
9829199 to
f163fa6
Compare
a777dfd to
3388dae
Compare
johncwaters
left a comment
johncwaters
left a comment
There was a problem hiding this comment.
Note
Automated review. Not written by a human.
- PR description: it lists
commandments.test.tsandsrc/shared/errors/__tests__/codes.test.tsas removed, but both files are only trimmed. Please update the description.
| }); | ||
| }); | ||
|
|
||
| describe('seam scan — routing reads live only in flags/', () => { |
There was a problem hiding this comment.
Note
Automated review. Not written by a human.
The seam scan is still deleted, and nothing replaces it. Every file it scanned (harness.ts, sequence.ts, models.ts, index.ts, orchestrator-runner.ts) still exists and none reads flags directly, so it would still pass. I couldn't find a lint rule or another test that keeps ctx.flags[...] reads and WIZARD_*_FLAG_KEY imports inside flags/. The block imports nothing from @programs, so could it stay?
| thinkingLevel: 'medium', | ||
| } as const; | ||
|
|
||
| describe('flag declarations', () => { |
There was a problem hiding this comment.
Note
Automated review. Not written by a human.
With these declaration checks gone, and the isolation test using a synthetic uncovered-program, nothing pins that wizard-orchestrator covers exactly posthog-integration or that self-driving covers exactly self-driving. Adding metrics to ORCHESTRATOR_SEQUENCE_ROUTE.programs would pass every test. These checks import only from flags/, so they fit the new layer rules. Please keep them.
3388dae to
989ddd1
Compare
f163fa6 to
2e19791
Compare
989ddd1 to
85e2539
Compare
2e19791 to
0de82f1
Compare
Generated-By: PostHog Desktop Task-Id: d14e92bb-6ee1-49b5-8502-39cb80079589
0de82f1 to
7bebc68
Compare
85e2539 to
1cda88f
Compare
edwinyjlim
left a comment
edwinyjlim
left a comment
There was a problem hiding this comment.
Review of #1387 at 7bebc68d. Comments inline.
Drafted with Claude Code from a review of the whole Release C stack, re-verified at the stack tip 2e31e438. Line numbers are at this PR's head.
| // Targeted assertions for the wizard_ask Path A translation rules. | ||
| // These are the rules a skill author depends on when leaving their prose | ||
| // unchanged — they need to keep working as the commandment list evolves. | ||
| describe('wizard_ask Path A rules', () => { |
There was a problem hiding this comment.
[P3] Eight commandments tests deleted although the commandments text did not change
src/agent/commandments.ts is byte-identical to main, and every phrase these tests matched is still there, so all eight would pass. The prompt snapshot tests do pin the full text, so this is not a total loss, but a blind snapshot update could now drop a Path A rule (wizard_ask named, no inline questions, batching, kebab-case ids, no re-asking) with nothing failing.
Suggested fix: Restore the two cases under "commandments by axis" and the whole "wizard_ask Path A rules" describe verbatim; only the joined-text constant needs to come back with them.
| }, | ||
| ); | ||
|
|
||
| it('says the handoff is nested', () => { |
There was a problem hiding this comment.
[P3] Orchestrator description and constant pins dropped though their subjects are unchanged
The complete-task description still says "one nested object" and names the handoff fields; TASK_NOTICE_TIMEOUT_MS is still exported. The "optional field" case is redundant and fine to drop.
Suggested fix: Restore the two description tests and the five-minute pin, or say in the body why they go.
3 participants
Phase 2 of Release C (#1320), PR 11 of 16, after #1386. Tests only. PRs 1 to 14 of this phase split one change into reviewable pieces and compile only as a stack: review them one by one, then land the whole phase together.
Problem
#1376 and #1378 move shared code out of the agent layer and change
RunConfig, so their tests must follow. Some tests locked only what is gone, such as thePROGRAM_BINDINGStable.Changes
Tests only: 38 files in
src/agentandsrc/shared.run-agent-linear-sdk.test.tscoversrunAgenton the linear SDK path (the run's prompt, abort, classified failures, progress order).console-log.test.tscoversprintAbortOutro.src/shared/__tests__/skill-install.test.tstakes the skill download cases fromwizard-tools.test.ts, zip-slip checks included.switchboard.test.tsandflags.test.tspass abindingin the context,agent-interface.test.tsandrun-agent-standalone.test.tsbuildRunConfigwithrouting, andgateway-session.test.tsfollows the per-run gateway setup.commandments.test.tsandsrc/shared/errors/__tests__/codes.test.ts.debug()cases, thePROGRAM_BINDINGScases, part ofhandoff-schema-parity.test.ts, and the settings-conflict case ofcredential-isolation-snapshot.test.ts, whichsettings-conflicts.test.tscovers. ThesessionPropertiesand ask-gate cases move to the programs tests in test: tests for the programs layer #1388.OutroKindfrom@shared/outro, and dropping the@uimocks now that@uiis gone.Test plan
pnpm typecheck0 errors,pnpm lint0 errors, 3,515 tests pass in 236 files,pnpm buildok.Created with PostHog Desktop