fix: stabilize workflow runtime and skill lifecycle surfaces#13
Merged
Conversation
- restore the requested anti-bot planning docs onto local main only - keep these files out of the merged remote main history Co-authored-by: Codex <noreply@openai.com>
- Preserve auth and anti-bot sessions through shared registry-backed fallback flows - Add pointer move/down/up/drag across CLI, tools, and /ops for manual challenge handling - Cover the new runtime contracts with fixture-backed provider, manager, and transport tests Co-authored-by: Codex <noreply@openai.com>
- Make the runtime audit self-sufficient with temp-harness lifecycle control and live probe scripts - Prevent ephemeral installs from mutating persistent autostart state during audit runs - Realign the shipped OpenDevBrowser skills and validators to the current runtime surface Co-authored-by: Codex <noreply@openai.com>
- Update public counts and inventories for the expanded CLI, tool, and /ops surfaces - Document the current challenge orchestration contract and legitimacy boundary - Keep README, architecture, CLI, and surface reference wording aligned with shipped behavior Co-authored-by: Codex <noreply@openai.com>
- add audit-domain metadata to the runtime matrix and orchestrator output - validate domain ownership and rerun contracts in the harness and asset checks - cover the expanded domain rollup behavior with focused runtime audit tests - sync the best-practices surface matrix to the broader audit inventory Co-authored-by: Codex <noreply@openai.com>
- avoid waiting for a fresh load event on fast invalid-credential transitions - expose the probe argument and branch helpers for direct regression coverage - lock the invalid-branch settle behavior with a focused script test Co-authored-by: Codex <noreply@openai.com>
- add the bounded challenge module, lane selection, verification, yield, and outcome recording flow - wire the new plane into direct browser, /ops, config, bootstrap, and provider fallback paths - preserve additive blocker metadata while exposing orchestration results without changing truth ownership Co-authored-by: Codex <noreply@openai.com>
- add focused unit coverage for evidence, interpretation, policy, lane selection, action loops, verification, yield, and outcomes - add runtime-handle integration coverage for direct browser and /ops challenge orchestration paths - extend provider fallback tests so preserved challenge state and orchestration metadata stay source-backed Co-authored-by: Codex <noreply@openai.com>
- update public surfaces to describe the shared bounded challenge plane and additive metadata - close the anti-bot challenge investigation with the current shipped ownership and legitimacy boundary - add the audited task-closure report with refreshed validation totals from the current gate run Co-authored-by: Codex <noreply@openai.com>
- Remove the anti-bot challenge audit, investigation, and planning docs requested for deletion - Keep unrelated local planning docs out of the commit - Verified lint, typecheck, build, extension build, and full test coverage gates before commit Co-authored-by: Codex <noreply@openai.com>
- Add the challenge automation override and desktop-boundary plan - Keep the plan aligned with live repo docs after the earlier anti-bot doc deletions - Verified lint, typecheck, build, extension build, and full test coverage gates before commit Co-authored-by: Codex <noreply@openai.com>
- Record the current generated date, branch, and baseline commit in the root agent guide - Close the last remaining tracked diff after the override-plan commit - Prior lint, typecheck, build, extension build, and full test coverage gates were green before this metadata-only commit Co-authored-by: Codex <noreply@openai.com>
- add run or session or config challenge automation precedence across managers, workflows, tools, and CLI commands - extend provider fallback recovery and telemetry so browser and helper decisions stay observable in runtime results - cover the new override and fallback behavior with targeted runtime and provider regression tests Co-authored-by: Codex <noreply@openai.com>
- preserve direct-run challenge orchestration telemetry across provider direct and shopping audit results - stop configured daemons cleanly without mutating autostart state during runtime audit cleanup - add focused regression coverage for the audit scripts and probe cleanup helpers Co-authored-by: Codex <noreply@openai.com>
- refresh README, CLI, surface, architecture, and privacy guidance around challenge automation controls and help ownership - update doc drift and help parity checks to match the shipped public surface and skill-pack metadata - remove the standalone challenge automation override plan document from docs Co-authored-by: Codex <noreply@openai.com>
- sync README and CLI skill fallback wording - extend docs drift enforcement for the March 23 closure surface Co-authored-by: Codex <noreply@openai.com>
- add deterministic validator fixture and resolver precedence coverage - strengthen per-pack validator execution checks and add the closure report Co-authored-by: Codex <noreply@openai.com>
- clear stale extension leases and surface relaunch-required recovery - bootstrap fresh attach targets when restricted tabs block reconnect - lock runtime-factory and extension recovery behavior with regressions Co-authored-by: Codex <noreply@openai.com>
- exclude zero-price offers from ranked output and report region enforcement truth - preserve auto-mode extension challenge boundaries in workflow metadata - refresh workflow, provider, and live-matrix regression coverage Co-authored-by: Codex <noreply@openai.com>
- capture shopping extension reliability findings and live validation evidence - record broader workflow defect analysis for shopping, research, and product lanes Co-authored-by: Codex <noreply@openai.com>
- defer playwright-core loading until chrome cookie bootstrap actually needs it - wrap corrupt or incomplete installs with explicit remediation guidance - cover lazy loading and memoized runtime import behavior Co-authored-by: Codex <noreply@openai.com>
- add interaction signals for popup, hold, and drag challenge surfaces - execute click-and-hold flows and follow verified target changes across attempts - strengthen challenge evidence, policy, orchestration, and verification regressions Co-authored-by: Codex <noreply@openai.com>
- capture current browser-help capability priorities from repo-backed evidence - record the challenge automation override plan and flag/workflow investigation Co-authored-by: Codex <noreply@openai.com>
- add a first-class review surface across tool, CLI, daemon, and extension ops paths - collapse provider runtime overrides into a canonical runtime policy resolver - make popup/ref/challenge execution paths consistent and cover them with regressions Co-authored-by: Codex <noreply@openai.com>
- update README and public surface docs for the review command and current channel inventory - refresh the flag/challenge investigation with the current validated seams and remaining live blocker - add the fallback compatibility investigation and align the best-practices skill audit artifacts Co-authored-by: Codex <noreply@openai.com>
- harden popup ownership, root attach recovery, and staged attach diagnostics - align browser, CLI, and relay routing with the new extension attach paths - extend regression coverage for extension routing, reconnect, and target use Co-authored-by: Codex <noreply@openai.com>
- disclose the webNavigation permission in privacy and store listing copy - add the focused popup attach probe runbook for direct launch failures Co-authored-by: Codex <noreply@openai.com>
- add design-canvas and bidirectional code-sync technical specs - record the Chrome DevTools MCP competitive inventory and refresh linked counts - document the extension-hosted canvas runtime guide and index the new specs Co-authored-by: Codex <noreply@openai.com>
- record the shopping-focused workflow autonomy investigation and root cause - extend the same analysis across research, product-video, and browser-surface packs Co-authored-by: Codex <noreply@openai.com>
- add missing scoped AGENTS guides for challenges, cli commands, and integrations - refresh root and module guide counts plus canvas and relay ownership notes - correct the historical landing metrics note to match current command and tool counts Co-authored-by: Codex <noreply@openai.com>
- replace inline build-copy shell code with a dedicated postbuild script - keep dist output normalization isolated from runtime and docs changes Co-authored-by: Codex <noreply@openai.com>
- record the current workflow inventory, surface map, and honest live-validation outcomes - refresh the autonomy execution ledger with the green closure proof for this branch Co-authored-by: Codex <noreply@openai.com>
- evict conflicting daemon owners before serve claims readiness - keep relay and ops state authoritative across attach and smoke flows - add regression coverage for browser, daemon, relay, and smoke paths Co-authored-by: Codex <noreply@openai.com>
- add source-driven workflow inventory and validation matrix scripts - align live and provider status contracts across shell-only and manual waits - cover workflow, provider, and youtube transcript probes with regression tests Co-authored-by: Codex <noreply@openai.com>
- update CLI, surface, and troubleshooting docs for the verified contract - expand docs drift coverage for the refreshed workflow inventory checks - keep documentation aligned with the current validated surface counts Co-authored-by: Codex <noreply@openai.com>
- add reusable runtime bundle and macro execution helpers - switch core bootstrap and workflow tools to the shared runtime path - cover init-override behavior in workflow runtime tests Co-authored-by: Codex <noreply@openai.com>
- add session-inspector across managed, ops, daemon, CLI, and tool surfaces - add dialog and upload command lanes plus screenshot and click timeout ergonomics - extend managed and extension ops coverage for dialogs, uploads, screenshots, and session review metadata Co-authored-by: Codex <noreply@openai.com>
- generate CLI and tool inventory from a shared public-surface source - drive help, docs, installer guidance, and bundled skill policy from the generated contract - add onboarding smoke coverage and refresh docs and skill assets around the best-practices entry lane Co-authored-by: Codex <noreply@openai.com>
- remove the remaining tool surface shim in favor of generated manifest exports - centralize alias-only and shadow-risk notes in shared onboarding metadata - surface those notes in help, installer output, local-only tools, and proof gates Co-authored-by: Codex <noreply@openai.com>
- replace stale tool-surface shim references with the generated manifest owners - document the one-cycle alias-only policy for research and shopping - call out the stale global best-practices shadow risk in docs and skill artifacts Co-authored-by: Codex <noreply@openai.com>
- sync canonical bundled skill installs refreshes and uninstall cleanup - align help metadata and local-only skill tooling with generated surface data - harden smoke harness startup and daemon status retries for closure validation Co-authored-by: Codex <noreply@openai.com>
- cover canonical skill sync refresh uninstall and legacy alias cleanup behavior - protect help onboarding smoke and daemon status retry expectations - keep loader filtering skill listing and workflow inventory parity explicit Co-authored-by: Codex <noreply@openai.com>
- update README CLI onboarding and surface reference copy to the canonical pack contract - remove retired alias packaging guidance from skill docs and roadmap notes - keep first-run guidance aligned with generated help and managed lifecycle proof Co-authored-by: Codex <noreply@openai.com>
- create the LaunchAgent log directory before plist write and bootstrap - add temp-home regression coverage for LaunchAgents and Logs ordering - keep the existing launchctl flow unchanged Co-authored-by: Codex <noreply@openai.com>
- fold duplicate smoke coverage into the canonical script test file - cover synthetic variant parsing and timeout diagnostics - keep the smoke script behavior unchanged while widening proof Co-authored-by: Codex <noreply@openai.com>
- ignore prompt-exports scratch outputs by default - remove the tracked popup oracle export artifact - keep local planning artifacts out of future cleanup diffs Co-authored-by: Codex <noreply@openai.com>
- drop superseded investigation packets from docs - keep the active parity divergence registry intact - isolate historical evidence cleanup from runtime changes Co-authored-by: Codex <noreply@openai.com>
- use viewport-relative clip coordinates for managed ref screenshots - add managed regressions for offscreen refs and stale snapshot guidance - characterize extension ref screenshot clip forwarding and stale-ref handling Co-authored-by: Codex <noreply@openai.com>
- forward 30s default transport budgets for snapshot, review, screenshot, page open, and target creation - preserve explicit timeout override precedence on the public CLI surface - cover default forwarding with focused CLI regression tests Co-authored-by: Codex <noreply@openai.com>
- await and verify fresh relay handshakes during reconnects - retry startup auto-pair mismatches instead of idling - require clean ext plus handshake readiness in CLI and session inspection - document the clean popup reconnect path for extension mode Co-authored-by: Codex <noreply@openai.com>
- harden popup and created-tab attach recovery in the extension runtime - retain router attach state across client resets and classify blocked attach errors Co-authored-by: Codex <noreply@openai.com>
- wait for ext=on and handshake=on before retrying recoverable ops failures - cover unknown-session and timeout recovery paths with relay-health checks Co-authored-by: Codex <noreply@openai.com>
- add discovery reports with winning-source, shadow, and issue metadata - preserve first-match-wins loading while surfacing audit visibility - cover cache-hit and default provenance branches in skill-loader tests Co-authored-by: Codex <noreply@openai.com>
- add a machine-checked continuity validator and register it in pack coverage - deepen login, form, and extraction validators with structural task-readiness checks - extend deterministic workflow-pack coverage to continuity and design-agent validators Co-authored-by: Codex <noreply@openai.com>
- rerun smoke live-regression failures against configured daemon state before defect attribution - surface provider coverage gaps explicitly and persist product review bundles for visual checks - expand best-practices runtime runbooks and audit matrices to match the enforced proof contract Co-authored-by: Codex <noreply@openai.com>
- update active docs and AGENTS guidance to current 64/57/59/35 surface counts - add nested annotate and macros AGENTS guides - document local-only artifact policy and config audit status Co-authored-by: Codex <noreply@openai.com>
- keep generic research inside public topical source families - harden shopping postprocess diagnostics for zero price, budget, and region-currency filtering - switch YouTube transcript defaults to public-first resolver behavior with truthful probe reporting - align workflow inventory and live matrix coverage with the validated capability lanes Co-authored-by: Codex <noreply@openai.com>
- forward shared workflow transport timeouts for research, shopping, and product-video commands - add validated transcript, research, and shopping lanes to help, onboarding, docs, and skill guidance - extend the best-practices router and CLI help tests to lock the reliable workflows Co-authored-by: Codex <noreply@openai.com>
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
Testing
Notes
region_authoritative=true