Repository navigation
fix(stage2): Android sign-in 401, iOS springboard relaunch, npm run cost dir - #145
Merged
Merged
Conversation
…ost dir - Android: the Sign Up password step no longer submits. mobile-mcp types submit as a literal "\n", so the account was registered with "ValidPassword1!\n" and the later sign-in got 401. - iOS: foregroundApp() traces launch errors instead of swallowing them and relaunches once if the screen still looks like the springboard. - npm run cost: default to tmp/trace/ of the invoking directory (INIT_CWD), where the agent writes its traces, not this repo's tmp/trace/. Run from an app workspace it silently reported a stale planner-only total. - README: measured VISUAL=2 time/cost from a passing Orange Baton run (7m 12s, $0.223) and a new validation report screenshot. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01UZJxUoubntzu7Ec7DPVfAr
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
I found these while running a paid-edition
NATIVEAPPTEMPLATE_VISUAL=2generation of the Orange Baton spec. Before the fixes, two runs failed Stage 2 on both mobile platforms. After them, the run passed: iOS 46/46, Android 48/48, 7m 12s, $0.223.scenarios/queue.ts): mobile-mcp sendssubmit: trueas a literal"\n". The Sign Up step therefore registered the password as"ValidPassword1!\n", which I confirmed in the Rails DB, and the later sign-in got 401. The password step now submits only on iOS.stage2-judge.ts): one run judged the springboard instead of the app, because a launch error was swallowed by a silentcatch {}.foregroundApp()now traces launch errors. If the screen still shows home-screen icons, it relaunches once. In the passing run the relaunch never fired, so the root cause is still unconfirmed; the next time it happens, the trace will show why.npm run costread the wrong directory (scripts/cost.sh): the agent writes traces totmp/trace/under the directory it ran in, but the script defaulted to this repo'stmp/trace/. Run from an app workspace, it printed a stale total with only a planner row. It now defaults to${INIT_CWD:-$PWD}/tmp/trace.Tests
Each fix has a regression test in
tests/smoke.test.ts. I watched each test fail against the original code, then pass with the fix:buildQueueScenario never submits a typed password on AndroidrunStage2Visual relaunches the app when the first launch leaves the iOS springboard upcost.sh with no argument reads tmp/trace/ of the directory npm run cost was invoked fromnpm test: 160/160 pass.npm run buildandtsc -p tsconfig.test.jsonare clean.npm --prefix <repo> run cost, run from the workspace, matches the hand-computed $0.2231.🤖 Generated with Claude Code
https://claude.ai/code/session_01UZJxUoubntzu7Ec7DPVfAr