github can1357/oh-my-pi v16.5.1
on GitHub

latest release: v16.5.2
5 hours ago

@oh-my-pi/pi-agent-core

Fixed

  • Fixed compatibility with Copilot gpt-5.6 models by correcting token escaping in compaction summaries.

@oh-my-pi/pi-ai

Added

  • Added Cursor OAuth and access-token usage reporting to omp usage via Cursor's account usage endpoint.

Fixed

  • Fixed OpenAI Responses content_filter terminal events being auto-retried as provider finish errors, ensuring content-filtered turns remain hard failures without triggering a retry loop.
  • Improved credential rotation on usage and account-quota failures to cycle through all eligible credentials instead of stopping early, while maintaining rate-limit backoffs and safety guards.
  • Fixed GLM tool call parsing to correctly handle and recover from missing or mistyped argument closers, preventing subsequent arguments from being swallowed.
  • Fixed Anthropic credential management and usage routing for users with multiple organizations under a single email. Credentials, OAuth refreshes, usage reports, and active sessions are now correctly partitioned and isolated by organization, preventing subscriptions from overwriting or merging with each other.
  • Fixed OpenAI and Codex response finalization to preserve streamed text when receiving empty content on completion. ([#5146])
  • Fixed OpenAI Chat Completions request parsing to correctly accept assistant tool-call replay messages with null content. ([#5121])
  • Fixed session-sticky OAuth credential mappings remaining active after credential changes, ensuring sessions correctly reselect accounts after login or logout. ([#4982])
  • Fixed concurrent reasoning summaries to ignore legacy streaming events under cutoff contracts.
  • Fixed Codex saved-reset redemption to apply to the selected OpenAI account in multi-account configurations. ([#5054])
  • Updated the OAuth completion page to instruct users to close the tab manually when the browser blocks automatic window closing. ([#4855])
  • Fixed Cursor max_mode requests to correctly send max-mode metadata on both model payload fields. ([#4797])
  • Fixed configuration discovery to support both nested and flat YAML formats for auth.broker.url and auth.broker.token keys. ([#4734])

@oh-my-pi/pi-catalog

Fixed

  • Fixed reasoning effort mapping for Z.ai GLM-5.2 on the Anthropic messages endpoint to correctly use the two-tier scale (high, max) and emit output_config.effort.
  • Fixed an issue where stale cached model limits would override updated static catalog limits after a catalog fingerprint mismatch.
  • Fixed Cursor discovery to correctly preserve GetUsableModels max-mode metadata for premium models and invalidate stale cache entries.

@oh-my-pi/pi-coding-agent

Changed

  • Enhanced Anthropic credential and usage management to support organization-scoped accounts, including displaying organization names in /usage, /logout, omp token --list, and OAuth login success messages, resolving active-account matching for shared organizations, and deduplicating identities during migration.

Fixed

  • Fixed compatibility of GNU-flavored shell builtins (such as stat, date, sed, mktemp, tail, find, base64, and ln) when invoked with macOS/BSD-style arguments.
  • Fixed subagent model and thinking level resolution to correctly respect the configured modelRoles.task selector instead of intermittently falling back to the parent session's model.
  • Fixed TUI rendering issues, including preventing macOS runtime diagnostics from painting into the viewport, bounding transcript retention in long sessions, and fixing scrollback repainting when collapsing history.
  • Fixed /tan and /fork clones failing to inherit or persist the parent session's prompt cache keys.
  • Fixed Python and JavaScript evaluation kernels suspending the CLI on subprocess foregrounding, deadlocking on non-serializable values, or losing in-flight subagent work during external aborts.
  • Fixed configured retry.fallbackChains failing to engage when encountering non-retryable provider errors.
  • Improved auto-compaction to automatically drop images and elide content when context is tight, and added persistent warning badges to the compaction divider when manual intervention is required.
  • Fixed the downshift plan nudge silently ending runs with no code written when the model answered with a text-only reply.
  • Fixed launch tool rendering and status reporting, including resolving contradictory readiness timeout messages and preventing backgrounded Bash blocks from continuing to repaint.
  • Fixed Advisor containment and timing issues, preventing hallucinated tool calls from contaminating later advice and ensuring late-arriving transcript deltas are coalesced before advisor calls.
  • Fixed omp update on npm-managed Windows installations to prevent downloaded release binaries from overwriting npm launchers.
  • Fixed --max-time duration values (e.g., 5s, 10m, 1h) being ignored instead of setting a session deadline.
  • Fixed omp plugin install --force failing with a dependency loop when replacing an existing pinned Git plugin source.
  • Fixed MCP tools receiving session image attachments as raw local:// URIs instead of resolving them to local filesystem paths.
  • Fixed Pyright LSP semantic requests hanging during startup.
  • Fixed Codex web search requests for GPT-5.6 Responses-Lite models.
  • Fixed custom model/provider configuration discovery to correctly load ~/.omp/agent/models.yaml when models.yml is absent.

@oh-my-pi/collab-web

Fixed

  • Fixed an issue in the live collaboration transcript where duplicate tool cards and a stale "thinking..." shimmer were rendered while a committed tool call was running.

@oh-my-pi/pi-tui

Fixed

  • Optimized the Markdown rendering cache to prevent large documents from indefinitely occupying cache slots, improving memory usage and performance (#4820).
  • Fixed viewport corruption on macOS caused by unmanaged stderr writes (such as libmalloc or framework diagnostics) while the terminal is active.
  • Fixed an issue where streamed diff code fences retained unhighlighted rows in native scrollback when long transient blocks left the viewport before finalization (#5126).
  • Fixed native Windows Terminal sessions failing to detect mid-run light/dark theme changes when Mode 2031 appearance notifications are unavailable (#5091).
  • Hid empty HTML comment separators in Markdown-rendered TUI output instead of displaying them literally (#4911).

@oh-my-pi/pi-utils

Added

  • Added terminal stderr guard utilities (suppressTerminalStderr and restoreTerminalStderr) to prevent macOS runtime diagnostics from corrupting TUI viewports while ensuring crash reports remain visible.

Fixed

  • Fixed an issue in Mermaid ASCII routing where unreachable edge attachment points caused unbounded pathfinder searches.

What's Changed

  • fix(coding-agent): suppress todo reminders for interactive questions by @roboomp in #5097
  • fix(tui): highlighted streamed diff scrollback rows by @roboomp in #5129
  • fix(coding-agent): support models.yaml custom config by @roboomp in #5147
  • fix(ai): preserve streamed Responses text on empty done items by @roboomp in #5149
  • fix(rpc): prevent rpc-ui extension UI response deadlock by @roboomp in #5156
  • fix(tui): display Cursor final answers after tool output by @roboomp in #5158
  • fix(tui): sanitize subagent progress output by @roboomp in #5162
  • fix(coding-agent): reject unknown continued session ids by @paralin in #5167
  • fix(session): recover interrupted session turns by @paralin in #5168
  • fix(ai): scope Anthropic credential identity by organization, not just email by @chan1103 in #5170
  • fix(coding-agent): load web search fallbacks lazily by @wolfiesch in #5182
  • fix(advisor): quarantine unknown tool responses by @roboomp in #5183
  • fix(coding-agent): preserve late advisor notes after terminal answers by @roboomp in #5186
  • fix(agent): escape Harmony compaction markers by @roboomp in #5190
  • fix(coding-agent): reject empty advisor stop completions by @roboomp in #5216
  • fix(coding-agent): accept Auto-Learn capture empty stops by @roboomp in #5217
  • fix(codex): preserve sequential reasoning summaries by @riverpilot in #5233
  • fix(tui): compact transcript retention without losing resize history by @RensTillmann in #5239
  • Fix automatic rollover across quota-limited accounts by @jeffscottward in #5247
  • fix(ai): route Z.AI API keys by usage by @roboomp in #5275
  • fix(coding-agent): respect role thinking for temporary model picks by @roboomp in #5297
  • fix(coding-agent): isolate eval runtimes from terminal by @masonc15 in #5327
  • fix(coding-agent): forward sessionId to getApiKey in subagent auth fallback (Closes #5325) by @djdembeck in #5377
  • fix(coding-agent): support GPT-5.6 Codex web search by @daandden in #5392
  • fix(auth-broker): resolve nested auth.broker.url/token yaml keys by @roboomp in #5406
  • fix(cli): show progress for print mode prompts by @roboomp in #4906
  • fix(tui): hide empty html comments by @roboomp in #4912
  • fix(coding-agent): skip autolearn capture after aborted turns by @metaphorics in #4924
  • fix(mcp): resolve local image paths for tool calls by @roboomp in #4950
  • fix(collab-web): dedup active transcript tools by @roboomp in #4952
  • fix(ai): clarify OAuth completion close copy by @roboomp in #4955
  • fix(catalog): preserve static limits on cache fingerprint mismatch by @roboomp in #4959
  • fix(plugin): handle pinned git source replacements by @roboomp in #4967
  • fix(prompting): allow magic keywords beside punctuation by @roboomp in #4968
  • fix(cursor): preserve max-mode flag by @roboomp in #4975
  • fix(ai): clear stale OAuth session stickies after credential changes by @roboomp in #4986
  • fix(eval): shield Python agent bridge aborts by @roboomp in #5015
  • fix(cli): parse max-time duration suffixes by @roboomp in #5044
  • fix(ai): apply codex reset to selected account by @roboomp in #5056
  • fix(cli): route npm self-updates through npm by @roboomp in #5057
  • fix(coding-agent): prevent memory URL writes from leaking to cwd by @roboomp in #5078
  • fix(agent): isolate memory root resolution by @roboomp in #5086
  • fix(advisor): reduce stale advisories via delta coalescing, WIP markers, and delivery annotation by @apoc in #5090
  • fix(agent): abort repeated empty yield results by @roboomp in #5099
  • fix(ai): stop retrying messages that were blocked by the provider by @belchetz in #5101
  • fix(catalog): classify Z.ai GLM-5.2 anthropic-messages endpoint as budget-effort with high/max by @oldschoola in #5119
  • fix(ai): accept null assistant chat content by @roboomp in #5123
  • fix(tui): suppress unmanaged macOS stderr writes corrupting the viewport by @korri123 in #5131
  • fix(coding-agent): flush throttled output tails by @wolfiesch in #5139
  • feat(ai): report Cursor account usage by @riverpilot in #5248
  • fix(lsp): send initial workspace configuration by @roboomp in #5277
  • fix(utils): bound Mermaid ASCII pathfinder by @roboomp in #5295

New Contributors

Full Changelog: v16.5.0...v16.5.1

Check out latest releases or
releases around can1357/oh-my-pi v16.5.1

Don't miss a new oh-my-pi release

NewReleases is sending notifications on new releases.

Get notifications