[renovate]

Note

This PR body was truncated due to platform limits.

This PR contains the following updates:

Package	Change	Age	Confidence
@oh-my-pi/pi-agent-core (source)	18.10.1 → 18.12.6
@oh-my-pi/pi-coding-agent (source)	18.10.1 → 18.12.6

---

Release Notes

can1357/oh-my-pi (@​oh-my-pi/pi-agent-core)

v18.12.6

Compare Source

@​oh-my-pi/pi-catalog

Added

• Added the prompt-cache-persistent catalog axis (isPromptCachePersistent) for providers whose own endpoint never expires prompt caches; logged-in OpenAI Codex declares it

Changed

• Updated the bundled Yolo-Auto catalog to its current models, qwen3.8-27b (now the default), yolo, yolo-small, and yolo-128k, and removed the retired deepseek-flash-v4 and qwen3.8-flash

@​oh-my-pi/pi-coding-agent

Added

• Added the _session/steering ACP extension (advertised as _meta.steering.supported), so clients such as agent-shell can steer a running turn; a steer sent while a permission dialog is open cancels that dialog, and _meta.steering.queue: "followUp" queues a follow-up instead (#​15885 by @​marcintustin)
• /mcp reauth <name> now works over ACP, RPC, and print mode (omp -p '/mcp reauth <name>'): it opens the browser, prints the authorization URL, reconnects the server in the running session, and print mode exits 1 when reauthorization fails (#​16046 by @​marcintustin)

Changed

• ACP session/cancel now also cancels prompts that are queued or still dispatching, and stops autonomous runs, instead of being ignored when no client turn is streaming (#​15885 by @​marcintustin)
• Idle compaction no longer runs on logged-in OpenAI Codex accounts, whose prompt cache never expires; compacting while idle only threw away a warm cache
• ACP sessions no longer fail to start when a client-declared MCP server cannot connect; the failure is reported in the session, and an HTTP/SSE server can be repaired with /mcp reauth (#​16046 by @​marcintustin)
• /mcp reauth, /mcp test, and /mcp unauth now pick the project entry when a server is defined in both project and user config, and skip project config when mcp.enableProjectConfig is off (#​16046 by @​marcintustin)
• /mcp reauth removes the server's stale Authorization header from its config when it writes the new OAuth binding (#​16046 by @​marcintustin)
• omp -p '/compact', /handoff, and /retry given as CLI arguments now run the command instead of sending the text to the model, and exit 1 when the command fails; slash text arriving on stdin or from @file is still sent to the model (#​16387 by @​marcintustin)
• Interactive startup no longer waits for extensions that register only tools, commands, flags, renderers, or event hooks (for example MCP bridge plugins): they load once the composer accepts input, and your first message waits for them instead. An extension is deferred only after one launch has confirmed it registers no providers, shortcuts, composer shapes, or file fallbacks. Launches that resume a session, pass --tools, or pass flags omp does not recognize still load every extension first (#​16684 by @​H4vC)

Fixed

• Fixed subagents whose role has an explicitly empty fallback chain falling back through the default chain when the default role uses the same model, including after a retry fallback, prewalk or context promotion briefly switched models or the subagent was revived (#​16548 by @​eggpeat)
• Fixed /usage and omp usage --json listing identity-less OAuth accounts such as Devin as missing usage beside their working reports (#​16549 by @​eggpeat)
• Fixed releasing or reaping a browser tab returning while Chromium still listed the page; the close now waits until Chromium reports the tab destroyed (#​16680 by @​H4vC)
• Fixed the advisor prompt forbidding all repeated advice, including real escalations: advisors may now re-raise a prior nit or concern as a blocker once new evidence makes it one (#​13887, #​16484 by @​pjh890321)
• Fixed Esc no longer interrupting in the TUI after a second MCP OAuth login replaced an unfinished first one (#​16046 by @​marcintustin)

@​oh-my-pi/pi-tui

Changed

• In Tern, centered sheets (/context, reports, keyboard shortcuts, BTW history, session info, the pause screen) fill the pane when it is narrower than 100 columns instead of floating with margins (#​16676 by @​H4vC)

Fixed

• Fixed the /agents and /models hubs in Tern squeezing the left sidebar as long lines (such as a streaming new-agent draft) widen the body pane

@​oh-my-pi/pi-utils

Added

• Added RpcConnection.requestCancellable() and Client.requestPermission(params, { signal }); client connections abort a handler and answer -32800 when the peer sends $/cancel_request, while agent connections keep running the request and return its real result (#​15885 by @​marcintustin)

What's Changed

• fix(coding-agent): clarify blocker escalation in advisor prompt by @​pjh890321 in #​16484
• feat(coding-agent): support ACP _session/steering for mid-turn client steering by @​marcintustin in #​15885
• feat(acp): support /mcp reauth outside the TUI by @​marcintustin in #​16046
• feat(tui): fill narrow Tern panes with centered sheets and pickers by @​H4vC in #​16676
• fix(coding-agent): confirm browser tab close and repair red CI tests by @​H4vC in #​16680
• perf(coding-agent): load startup-safe extensions after the composer accepts input by @​H4vC in #​16684
• fix(coding-agent): keep explicitly empty subagent fallback chains empty by @​eggpeat in #​16548
• fix(coding-agent): stop listing identity-less OAuth accounts as missing usage by @​eggpeat in #​16549

New Contributors

• @​pjh890321 made their first contribution in #​16484
• @​marcintustin made their first contribution in #​15885

Full Changelog: can1357/oh-my-pi@v18.12.5...v18.12.6

v18.12.5

Compare Source

@​oh-my-pi/pi-ai

Added

• Added OpenAI 12-hour prompt caching for GPT-5.6+ on the Responses API: cacheRetention: "long" now requests prompt_cache_options.ttl: "12h" (instead of the ignored legacy prompt_cache_retention: "24h"), and promptCache.ttl accepts "12h"

Fixed

• Fixed every OpenAI Codex request failing with HTTP 404 {"detail":"Not Found"} on personal ChatGPT accounts whose token is not region-pinned (#​16493)

@​oh-my-pi/pi-catalog

Fixed

• Fixed Moonshot kimi-k3 pricing cache writes at $0; usage cost now includes cache writes at Moonshot's $3/M 5-minute rate (#​16508)

@​oh-my-pi/pi-coding-agent

Added

• providers.cacheRetention: long now keeps OpenAI GPT-5.6+ prompt caches for 12 hours (Responses API; cache writes cost 2.5× input instead of 1.25×)

Fixed

• Fixed subagents spawning at a caller-supplied effort (e.g. hi → max) while task.enableEffort is off; the field is now ignored and the agent's configured thinking level (including :auto) applies (#​16402)
• Fixed omp update --check failing against JFrog Artifactory registries that add empty dist-tags to a tagged manifest; the update check now uses the registry's curated latest version (#​16375)
• Fixed <path>:conflicts on a file past the 10 MiB scan cap reporting no conflicts with no hint: it now names the line the scan stopped at and the ranged read that registers the blocks after it (#​16269)
• Fixed the ts-set-map rule stalling tool-call streaming for seconds on a long single-line Map literal, and missing Map entries written ["key", value] as const (#​16242)
• Fixed the ts-import-type rule missing inline type imports whose member name is non-ASCII, such as import("./units").温度 (#​16242)

@​oh-my-pi/pi-natives

Fixed

• Fixed a macOS background action leaving its target in front of your app after you clicked another of your windows, when placing the click took longer than the click itself (#​16217)

What's Changed

• fix(update): handled empty Artifactory dist-tags on tagged manifests by @​roboomp in #​16379
• fix(task): ignored caller effort while task.enableEffort is off by @​roboomp in #​16406
• fix(catalog): stopped declaring no_constraint codex residency by @​roboomp in #​16501
• fix(catalog): priced moonshot kimi-k3 cache writes at the 5min rate by @​roboomp in #​16529

Full Changelog: can1357/oh-my-pi@v18.12.4...v18.12.5

v18.12.4

Compare Source

Fixed

• Enabled native compaction for supported Bedrock Claude models (#​16343 by @​anatoli-tsinovoy)

v18.12.3

Compare Source

Added

• AgentTool gained the optional approvalCacheTargets hook, which the always-ask-cached approval mode uses to check and record per-file grants (#​16049 by @​Peribunt)

Changed

• Tools with lenientArgValidation now receive unknown root arguments that their closed schema would otherwise drop, so they can reject them (#​16317)

v18.12.2

Compare Source

Changed

• The failed-tool-call correction sent to DeepSeek models now also covers chat-template tool-call tokens written out as text, not only malformed DSML (#​16216 by @​H4vC)

Fixed

• Fixed Ollama chat history estimates counting stripped native thinking and missing foreign thinking replayed as text, which skewed compaction after model switches (#​16164)

v18.12.1

Compare Source

@​oh-my-pi/pi-ai

Fixed

• Fixed every request failing behind a TLS-inspecting proxy that refuses compressed uploads: a 400/403/415 whose body is not JSON, or the Anthropic API's "request body is not valid JSON" after a proxy drops Content-Encoding, now resends the body uncompressed and stops compressing for that host (#​16108 by @​alphastorm)

@​oh-my-pi/pi-catalog

Fixed

• Anthropic's request-encoding-rejection now also matches the API's "request body is not valid JSON" 400, which a gzip body gets when a proxy drops its Content-Encoding header (#​16108 by @​alphastorm)

@​oh-my-pi/pi-coding-agent

Breaking Changes

• Removed display() from eval in both Python and JavaScript; print (and console.log in JavaScript) covers it

Added

• Added advisor.continueSeverity (nit, concern, blocker) to choose the minimum advisor note severity that wakes the agent after it finished with a final answer; lower-severity notes stay as a card at the end of the transcript
• Printed DataFrames/Series render as tables, large dicts/lists and objects as JSON trees, and Markdown objects as Markdown for the user, while the model sees the printed text
• Printing a figure or image in eval attaches the image to the result

Changed

• An advisor concern raised after the agent's final answer now continues the run by default instead of waiting at the end of the transcript; set advisor.continueSeverity to blocker for the previous behavior

Fixed

• Fixed context-file @ imports failing to load paths containing backslash-escaped spaces (#​16090)

@​oh-my-pi/pi-natives

Fixed

• Fixed native memory growing without bound, and eventually crashing the process, when files in a downloaded (WASM) grammar language such as C# were parsed repeatedly (#​16067)

@​oh-my-pi/pi-tui

Breaking Changes

• numberedCode, readContentCode and nativeDiff take the file's path (path) instead of a language, and describeGroupedOutput drops its lang option: Tern picks the grammar of file contents from their path

Changed

• Advisor notes in Tern read as a callout tinted by their worst severity: its icon and a tally (Concern, 1 blocker · 2 concerns) head the card with the advisor's name at the end, and several notes each lead with their own severity icon

Fixed

• Fixed images rendering blank inside tmux on Kitty-capable terminals when tmux's allow-passthrough is off (the default); omp now falls back to SIXEL or the text placeholder instead of sending Kitty graphics tmux drops
• Fixed Tern drawing tool cards' file contents (write, read, edit, grep, find, ast-grep, ast-edit) plain or in the wrong language when omp did not know the file's extension (.ptx) or mapped it to a neighbor (.cu as C++); omp now sends the file's path and Tern picks the grammar

What's Changed

• fix(ai): resend uncompressed when a proxy rejects a compressed request body by @​alphastorm in #​16108
• fix(tui): answered allow-passthrough in fake tmux client fixture by @​roboomp in #​16112
• fix(ast): retired per-thread wasm parser store after a byte budget by @​roboomp in #​16078
• fix(coding-agent): support escaped spaces in context imports by @​roboomp in #​16091

Full Changelog: can1357/oh-my-pi@v18.12.0...v18.12.1

v18.12.0

Compare Source

Fixed

• Fixed Cursor turns failing with PROVIDER_ERROR when a follow-up (a finished background job, a late diagnostic, a reminder) arrived while the previous turn was still streaming (#​15821 by @​robfeldmann)

v18.11.3

Compare Source

Added

• prepareQueuedMessages receives which queue (steering or followUp) a batch came from (#​16026)

v18.11.2

Compare Source

Added

• agent_start carries resumedToolCallIds when a continued run first executes the calls of an unpaired assistant tail

Fixed

• Cancelling an ask after a background job finishes now reports the user cancellation instead of telling the model to retry a skipped tool (#​15985)

v18.11.1

Compare Source

Added

• Tokenizer.countLatestTurn: counts a message as the transcript's latest assistant turn, which an Anthropic Messages replay keeps byte-for-byte (#​15791 by @​will-bogusz)
• Agent.prioritizeSteering: moves given pending steering messages ahead of older ones so one-at-a-time delivery takes them first (#​15642)

Fixed

• Fixed a Cursor tool call re-sent by the server persisting two results under one id; the re-run's result replaces the first (#​15749)
• Fixed compaction keeping more recent history than keepRecentTokens when the latest Claude turn on the Anthropic Messages API ended its tool calls on stop or length: the replay keeps that turn's signed reasoning, but the cut point counted only its visible blocks (#​15791 by @​will-bogusz)

v18.11.0

Compare Source

Breaking Changes

• Removed ApprovalDetailsOptions; formatApprovalDetails takes only args again, since the compact approval prompt no longer shows tool details (#​15642 by @​H4vC)

Added

• AgentTool.delegationGuidance: instructions a same-name extension replacement inherits in its description when it delegates to the native tool (#​15477 by @​iliaal)
• CompactionPreparation.lastSummarizedEntryId and CompactionResult.providerReplayThroughEntryId: a remote compaction that could only fit the history before the kept tail reports the last entry its native history covers; later entries replay normally (#​15712 by @​H4vC)

Fixed

• Fixed native (OpenAI/Codex) remote compaction refusing or truncating sessions that fit the window: its preflight counted the JSON-escaped request, inflating quote-, backslash- and newline-heavy tool output such as disassembly by ~10–20%; it now counts text as the model reads it (#​15712 by @​H4vC)
• Fixed OpenAI V1/V2 remote compaction failing outright when the history, kept tail included, exceeds the context window (a large final response or tool output after the last check); it retries on the history before the kept tail (#​15712 by @​H4vC)
• Fixed fitted output caps overflowing the context window on strict hosts (llama.cpp-family) after turns with many parallel tool results; chat-template framing is now charged per locally counted message (#​15699)
• Fixed leaked DeepSeek DSML markup reaching saved history from reasoning blocks and from turns that also made a valid tool call (#​15485)

v18.10.4

Compare Source

Added

• AgentTool.formatApprovalDetails receives ApprovalDetailsOptions; compact: true asks for untruncated bodies so a compact approval prompt can count their lines (#​15550 by @​H4vC)

Changed

• Before-model-call hooks now receive the loop context as a second argument, and buildProviderContext receives the summarizer model (#​15436 by @​will-bogusz)

Fixed

• Fixed a Claude compaction or branch summarizer receiving assistant reasoning from a snapcompact archive rendered under another model, which Claude refuses (reasoning_extraction) (#​15436 by @​will-bogusz)

v18.10.3

Compare Source

Breaking Changes

• Removed estimateImageContentTokens() from @oh-my-pi/pi-agent-core/image-tokens; use Tokenizer.imageTokens(size, detail), which prices with the reading model's catalog image rule (#​15446 by @​will-bogusz)
• applyShakeRegion() and applyShakeRegions() take the replacement's token count (replacementTokens); a block rewrite of an assistant turn adds the tokens it removes to elidedOutputTokens, so a replayed Claude turn's count stays lowered after a reload (#​15446 by @​will-bogusz)

Fixed

• Fixed Claude context counts reading Opus 5.5 requests about 55% over their billed input: a replayed thinking turn now counts at the output size Claude billed for it, not its signature's bytes (#​15446 by @​will-bogusz)
• Fixed screenshots and other non-frame images being counted at the OpenAI rate for every model instead of the reading model's own; Claude billed them about 16% higher (#​15446 by @​will-bogusz)
• Fixed OpenAI Responses and Codex context estimates leaving out replayed reasoning: each turn whose encrypted reasoning actually replays now counts its reported reasoning tokens (its ciphertext when the backend reports none), and none while a cold session rebuilds it out of the request, so compaction and snapcompact sizing match billing; branch summaries no longer budget for encrypted reasoning they never send (#​15447 by @​will-bogusz)

v18.10.2

Compare Source

Added

• Added AgentLoopConfig.onDetachedToolUpdate: a tool's onUpdate frames that arrive after its run's event stream ended (a returned call whose background job keeps reporting) reach it instead of being dropped; Agent forwards them to its subscribers as tool_execution_update events

can1357/oh-my-pi (@​oh-my-pi/pi-coding-agent)

v18.12.6

Compare Source

Added

• Added the _session/steering ACP extension (advertised as _meta.steering.supported), so clients such as agent-shell can steer a running turn; a steer sent while a permission dialog is open cancels that dialog, and _meta.steering.queue: "followUp" queues a follow-up instead (#​15885 by @​marcintustin)
• /mcp reauth <name> now works over ACP, RPC, and print mode (omp -p '/mcp reauth <name>'): it opens the browser, prints the authorization URL, reconnects the server in the running session, and print mode exits 1 when reauthorization fails (#​16046 by @​marcintustin)

Changed

• ACP session/cancel now also cancels prompts that are queued or still dispatching, and stops autonomous runs, instead of being ignored when no client turn is streaming (#​15885 by @​marcintustin)
• Idle compaction no longer runs on logged-in OpenAI Codex accounts, whose prompt cache never expires; compacting while idle only threw away a warm cache
• ACP sessions no longer fail to start when a client-declared MCP server cannot connect; the failure is reported in the session, and an HTTP/SSE server can be repaired with /mcp reauth (#​16046 by @​marcintustin)
• /mcp reauth, /mcp test, and /mcp unauth now pick the project entry when a server is defined in both project and user config, and skip project config when mcp.enableProjectConfig is off (#​16046 by @​marcintustin)
• /mcp reauth removes the server's stale Authorization header from its config when it writes the new OAuth binding (#​16046 by @​marcintustin)
• omp -p '/compact', /handoff, and /retry given as CLI arguments now run the command instead of sending the text to the model, and exit 1 when the command fails; slash text arriving on stdin or from @file is still sent to the model (#​16387 by @​marcintustin)
• Interactive startup no longer waits for extensions that register only tools, commands, flags, renderers, or event hooks (for example MCP bridge plugins): they load once the composer accepts input, and your first message waits for them instead. An extension is deferred only after one launch has confirmed it registers no providers, shortcuts, composer shapes, or file fallbacks. Launches that resume a session, pass --tools, or pass flags omp does not recognize still load every extension first (#​16684 by @​H4vC)

Fixed

• Fixed subagents whose role has an explicitly empty fallback chain falling back through the default chain when the default role uses the same model, including after a retry fallback, prewalk or context promotion briefly switched models or the subagent was revived (#​16548 by @​eggpeat)
• Fixed /usage and omp usage --json listing identity-less OAuth accounts such as Devin as missing usage beside their working reports (#​16549 by @​eggpeat)
• Fixed releasing or reaping a browser tab returning while Chromium still listed the page; the close now waits until Chromium reports the tab destroyed (#​16680 by @​H4vC)
• Fixed the advisor prompt forbidding all repeated advice, including real escalations: advisors may now re-raise a prior nit or concern as a blocker once new evidence makes it one (#​13887, #​16484 by @​pjh890321)
• Fixed Esc no longer interrupting in the TUI after a second MCP OAuth login replaced an unfinished first one (#​16046 by @​marcintustin)

v18.12.5

Compare Source

Added

• providers.cacheRetention: long now keeps OpenAI GPT-5.6+ prompt caches for 12 hours (Responses API; cache writes cost 2.5× input instead of 1.25×)

Fixed

• Fixed subagents spawning at a caller-supplied effort (e.g. hi → max) while task.enableEffort is off; the field is now ignored and the agent's configured thinking level (including :auto) applies (#​16402)
• Fixed omp update --check failing against JFrog Artifactory registries that add empty dist-tags to a tagged manifest; the update check now uses the registry's curated latest version (#​16375)
• Fixed <path>:conflicts on a file past the 10 MiB scan cap reporting no conflicts with no hint: it now names the line the scan stopped at and the ranged read that registers the blocks after it (#​16269)
• Fixed the ts-set-map rule stalling tool-call streaming for seconds on a long single-line Map literal, and missing Map entries written ["key", value] as const (#​16242)
• Fixed the ts-import-type rule missing inline type imports whose member name is non-ASCII, such as import("./units").温度 (#​16242)

v18.12.4

Compare Source

Added

• bash tool commands now see PI_SESSION_ID and PI_SESSION_FILE, so a script can find the running session's ID and transcript path (#​16442 by @​johnrichardrinehart and @​will-bogusz)
• Added GitLab merge request support to /review and /annotate code-review: pass an MR URL, pick one mentioned in the conversation, or choose "Review a specific PR/MR" in a GitLab checkout to list open MRs (requires glab); large MRs link per-file diffs through the new mr://<host>/<group>/<project>/<iid>/diff URLs (#​16361 by @​Shadorain)
• Added MCP tool progress: tools/call requests ask the server for notifications/progress when the progress can be shown, and updates appear in the running tool card and as an in-place mcp_progress status line in Eval cells, including after auto-backgrounding; request deadlines and Eval timeouts are unchanged (#​16431 by @​alphastorm)
• Added image attachments to single and batch eval judge()/judge_batch() calls (local paths or embedded images); image judgments skip candidates without image support and fail explicitly instead of dropping images (#​15471 by @​anatoli-tsinovoy)
• Added a job tree to /jobs, /jobs full and the jobs pill's sheet, in the terminal UI and over ACP: subagents' background jobs now appear nested under the task that runs them, with the busiest branches listed first; only your own jobs offer Cancel (#​15854 by @​H4vC)
• Added ctx.ui.getEditorComponent() so extensions can wrap the active custom editor factory instead of replacing it (#​8655 by @​jovi-tsx)
• omp auth-gateway serve and check can be limited to the providers and API keys the account-pool file allows, using "*": false and "apiKeys": false (#​14235 by @​will-bogusz)
• omp usage shows an account's reclaimAbovePct next to its priority and reserve (#​14237 by @​will-bogusz)

Changed

• An agent that messages another agent can now wait for the reply instead of erroring; the wait ends early, without an error, once the agent asked finishes its turn without replying or is itself waiting on the caller's result (#​16445 by @​will-bogusz)
• Built-in style rule reminders no longer tell the model a user-defined rule matched and it MUST comply; they ask it to check its code against the rule and its exceptions, so models rewrite code a built-in rule allows less often (#​16420 by @​will-bogusz)
• read of a SQLite ?q=SELECT … query now accepts &limit= and &offset= and pages the query's rows, as table reads and JSON ?q= reads already do, instead of rejecting the read (#​16444 by @​will-bogusz)
• Agents using the browser tool are steered to wait for the text, URL or element they need after a click or navigation instead of sleeping a fixed time, which can finish steps sooner on pages that load content after a variable delay (#​16404 by @​will-bogusz)
• The pinned Subagents list shows how many agents are running and the agent hub key, and previews an agent's assignment at full width instead of the first 40 characters of its prompt template (#​16486 by @​H4vC)

Fixed

• Fixed auto-QA reports whose tool is a URI (xd://tui: …) being filed under the wrong tool name; a one-line report without : after its tool name is now refused with the format hint (#​16278 by @​will-bogusz)
• Fixed agents guessing conflict://<N> ids for merge conflicts a ranged read didn't show: the warning and the not-found error now point to <path>:conflicts to register them (#​16269 by @​will-bogusz)
• Fixed agents opening your own Chrome through the relay (with browser.relay on) when they asked the browser tool for Chromium with app.tern: false; the tool now points them to app.relay: false (#​16294 by @​will-bogusz)
• Fixed %pip install in Python eval failing with "No module named pip" when the kernel's interpreter is a venv without pip, such as one made by uv venv: it now installs pip into that venv first, with ensurepip or, where that is unavailable, uv (#​16300 by @​will-bogusz)
• Reduced HTTP 413 failures on Factory Droid once a session's screenshots and pasted images pass Factory's 4.5 MB request limit: older inline images are now dropped to keep them under 3.5 MB, while the newest message's and newest tool result's images stay when they alone fit the limit; a custom gateway baseUrl is left alone (#​16288 by @​will-bogusz)
• Fixed hashline edits under an older tag overwriting the wrong line when their line numbers came from grep -n, sed -n or a :raw read: a shifted edit whose body matches the current line at those numbers, or that breaks the file's syntax, is now rejected with the current tag and lines (#​16383 by @​will-bogusz)
• Fixed PI_SESSION_FILE in eval cells being a relative path when omp was started with a relative --session-dir (or PI_CODING_AGENT_SESSION_DIR), so it no longer worked once code changed directory (#​16442 by @​will-bogusz)
• Fixed waking a parked subagent re-sending its whole conversation without the prompt cache whenever other agents had started, finished or parked meanwhile (#​15523) (#​16369 by @​will-bogusz)
• Fixed edit accepting changes to lines nothing showed after a stale or unknown tag: a resend under the current tag from the rejection now passes the seen-line check only on the rows that rejection printed (or a retained read showed), so a line seen only through bash is no longer overwritten blind (#​16422 by @​will-bogusz)
• Fixed plugin settings marked secret being readable by other local users in omp-plugins.lock.json; the lockfile is now written owner-only (0600) (#​12999, #​16368 by @​jonasboos)
• Fixed agents misreading computer.windows() results on their first call (#​16299 by @​will-bogusz)
• Fixed GitLab merge request reads rendering [object Promise] instead of the description (#​16477 by @​adi-IL)
• Fixed /restart losing the active named profile and failing to resume its session (#​15742 by @​sapran)
• Fixed aborted and timed-out stdio MCP requests continuing to run on the server: OMP now sends notifications/cancelled with the original request id, once and never for initialize, after settling the call and without waiting on the pipe (#​16430 by @​alphastorm)
• Fixed the agent restarting on its own after you stopped it with Esc whenever a background subagent or job finished; the result is recorded in the transcript and the agent waits for your next message
• Fixed Mnemopi memory completions ignoring the memory role's thinking suffix, including one inherited from tiny or smol when modelRoles.memory is unset; each retry.fallbackChains.memory entry sends its own level, and :off disables reasoning (#​16378 by @​alphastorm)
• Fixed MCP tools running a call twice when the server connection dropped mid-request: a stdio request already written to the server, or an HTTP or legacy SSE request whose connection reset or closed after it was sent, now fails as non-retryable with a diagnostic saying it may have run, instead of being replayed after a reconnect; requests that never left OMP, including refused connections, still reconnect and retry once (#​16429 by @​alphastorm)
• Fixed /move path completion jumping into a subfolder on Tab: .., ., and ~ now complete to ../, ./, and ~/ before listing their children, a trailing \ lists a folder on Windows, absolute Windows paths (C:/…) are no longer rewritten as relative paths, and the project directory is offered as ./ instead of /
• Fixed edit blaming boundary indentation when a one-row replacement only fails because its body ends with a trailing comma (e.g. the last element of a JSON array); the error now names the comma (#​16423 by @​will-bogusz)
• Fixed a timed-out browser selector action dropping its match count without a word when omp could not count the matches; the error now says why (#​16389 by @​will-bogusz)
• Fixed every session in a project without .vscode/mcp.json logging a "[VS Code] Failed to read …/.vscode/mcp.json" warning; a missing file is now skipped silently, while an unreadable or unparsable one still warns (#​16434 by @​will-bogusz)
• Fixed agents opening pages in your own signed-in Chrome when browser.relay is on and you asked for a throwaway browser (#​16390 by @​will-bogusz)
• Fixed the built-in ts-no-tiny-functions and ts-promise-with-resolvers rules reminding you about code they allow: exported functions, public class and object members, type guards, if/try/catch blocks, and never-settling new Promise(() => {}) (#​16419 by @​will-bogusz)
• Fixed read losing the last lines of a comma list that pairs ranges with a tail (file:1-30,-30, file:1-30,:-30), and erroring on the same list for artifact:// and other internal URLs; both parts are now read (#​16435 by @​will-bogusz)
• Fixed edit rejecting a placeholder tag (#AUTO, #????), a missing tag, or a header with no ops without naming the file's current tag; these now get the same reply as an unknown tag, so the retry can use that tag instead of a read (#​16425 by @​will-bogusz)
• Fixed read line ranges and tails (:9-11, :-50) of macOS .sample.txt reports and .cpuprofile files returning lines of the generated profile summary instead of the file, so edits from them were rejected as stale; the summary now appears only on a whole-file read, without line numbers (#​16397 by @​will-bogusz)
• Models whose provider strips images from requests, such as DeepSeek V4 Flash on Venice or NanoGPT, now get text compaction and text-only image handling like text-only models; snapcompact archived onto frames that reached them as [image omitted] placeholders, and history images were still processed and uploaded before being dropped (#​16436 by @​will-bogusz)
• Line range reads of ssh://, URLs and other in-memory resources no longer return only the line above the range when that line fills the 50KB budget; the page now starts at the requested line, and byte-limit notices state the 50KB limit instead of the bytes shown (#​16391 by @​will-bogusz)
• Fixed Auto-QA grievances sharing your home directory and login name with the qa.omp.sh collector; paths under your home are sent as ~/…, and the login name inside labels and paths as <user>, including reports recorded before this fix (#​16409 by @​will-bogusz)
• Fixed the computer tool's control.acquire prompt, left unanswered, failing the whole call with "Computer code execution timed out"; the prompt now expires shortly before the timeout and the call returns reason: "unanswered" (#​16388 by @​will-bogusz)
• Fixed reads whose optional leading context line exceeded the byte budget returning only the oversized-line notice or a preview of that context line and hiding the requested lines; the read now retries from the requested start without leading context (#​16485 by @​adi-IL).
• Fixed read path:N:raw (and :raw:N, URLs included) returning a page from line N instead of the one line N; grep on internal and web URLs now filters :N:raw to that line too, and :N-:raw still runs on from N (#​16438 by @​will-bogusz)
• Fixed Anthropic web search answers breaking into separate paragraphs at every citation, with sentences split mid-way and lines starting with . (#​16396 by @​a-glapinski)
• Fixed custom tools that import an @oh-my-pi/* package (such as @oh-my-pi/pi-natives or @oh-my-pi/pi-utils) failing to load with "Cannot find package" in the installed omp binary; they now get the host's copy, as extensions already did (#​16421 by @​will-bogusz)
• Fixed resuming a session selecting an abandoned branch after a late off-branch entry such as a background bash result; off-branch entries now record the active leaf they interrupted and replay restores it (#​16483 by @​adi-IL).
• Fixed hashline REM deleting and MV moving a file under an invented or stale tag; a bare file op now needs the file's current tag, like a line edit (#​16381 by @​will-bogusz)
• Fixed write deleting the leading numbers of every line in content such as hex dumps, disassembly and timestamped logs; it now removes copied read line numbers only when they count up by one, as read and grep print them (#​16441 by @​will-bogusz)
• Fixed auto thinking chosen for the DEFAULT role in the /model hub reverting to the model's default effort (e.g. max on Z.AI glm-5.3) on the next launch (#​16370, #​16398 by @​alvadorn)
• Fixed Google Antigravity web search failing with false 429 quota-exhausted errors (#​11849 by @​jotafurtado)
• Fixed omp exiting to the shell when the model picker (Alt+P) or model hub opened after compatibility extensions loaded (#​12293, #​12653 by @​wolfiesch)
• Fixed todo lists set over RPC set_todos disappearing when the session is reloaded, resumed, or compacted; RpcClient.setTodos() sends the command (#​12590 by @​anthonyrisinger)
• Fixed batch task calls silently running the default agent when a specialist such as reviewer, scout, or a custom agent was set at the top level instead of in tasks[].agent; the call is now rejected with that guidance (#​12610 by @​dbc-hbin)
• Fixed sessions started before an omp upgrade losing word completion and restarting the shared word-completion daemon every 30 seconds, which reloaded its model for every session on the new version; they now use the newer daemon (#​15610) (#​16380 by @​will-bogusz)
• Fixed supported Bedrock Claude models falling back to soft instead of remote compaction (#​16343 by @​anatoli-tsinovoy)
• Fixed browser tab.run(fn, [url]) and browser.open("https://…") failing with a bare "expects an options object"; the errors now say to pass { args: [...] } and browser.open({ url, name }) ([#​16236](https://redirect.github.com/can135

[!IMPORTANT]
✂ PR body was truncated to here.

---

Configuration

📅 Schedule: (in timezone Asia/Jerusalem)

• Branch creation
  • "after 10pm every weekday,every weekend,before 5am every weekday"
• Automerge
  • At any time (no schedule defined)

🚦 Automerge: Enabled.

♻ Rebasing: Whenever PR is behind base branch, or you tick the rebase/retry checkbox.

🔕 Ignore: Close this PR and you won't be reminded about these updates again.

---

• If you want to rebase/retry this PR, check this box

---

This PR was generated by Mend Renovate. View the repository job log.

[renovate]

⚠️ Artifact update problem

Renovate failed to update an artifact related to this branch. You probably do not want to merge this PR as-is.

♻ Renovate will retry this branch, including artifacts, only when one of the following happens:

• any of the package files in this branch needs updating, or
• the branch becomes conflicted, or
• you click the rebase/retry checkbox if found above, or
• you rename this PR's title to start with "rebase!" to trigger it manually

The artifact failure details are included below:

File name: paseo-omp/npm-shrinkwrap.json

npm warn Unknown env config "store". This will error in a future major version of npm. See `npm help npmrc` for supported config options.
npm error code EUNSUPPORTEDPROTOCOL
npm error Unsupported URL Type "catalog:": catalog:
npm error A complete log of this run can be found in: /runner/cache/others/npm/_logs/2026-10-24T21_56_48_816Z-debug-0.log

[coderabbitai]

Note

Reviews paused

It looks like this branch is under active development. To avoid overwhelming you with review comments due to an influx of new commits, CodeRabbit has automatically paused this review. You can configure this behavior by changing the reviews.auto_review.auto_pause_after_reviewed_commits setting.

• Last reviewed commit: bd401b2

Use the following commands to manage reviews:

• @coderabbitai resume to resume automatic reviews.
• @coderabbitai review to trigger a single review.

Use the checkboxes below for quick actions:

• ▶️ Resume reviews
• 🔍 Trigger review

No actionable comments were generated in the recent review.

Recent review info

Run configuration

• Source mode: Auto
• Repository: Unavailable
• Branch: Unavailable
• File: No configuration file from this repository applied
• Commit: Unavailable
• Layers: Organization UI
• Review profile: ASSERTIVE
• Plan: Advanced
• Run ID: f28c2324-2a81-4d51-8134-36afd593ad0d

Commits

Reviewing files that changed from the base of the PR and between f482e1d and bd401b2.

Files selected for processing: 1

• paseo-omp/package.json

Included review availability: This review used your included allowance. Your plan provides up to 10 included reviews per hour; 9 remain after this review.

Walkthrough

The paseo-omp manifest updates @oh-my-pi/pi-agent-core and @oh-my-pi/pi-coding-agent from version 18.10.1 to 18.12.6.

Changes: Development Dependency Update

Layer / File(s)	Summary
Update development dependency versions paseo-omp/package.json	The manifest updates both development dependencies to version 18.12.6.

Priority: Low

Estimated code review effort: 1 (Trivial) | ~3 minutes

Change: Other

Possibly related PRs

• omercnet/paseo-plugins#402: This PR adds @oh-my-pi/pi-coding-agent@18.12.6 to the Sandbox runtime and tests the omp CLI at that version.
• omercnet/paseo-plugins#399: This PR releases the Sandbox runtime that includes @oh-my-pi/pi-coding-agent@18.12.6.
• omercnet/paseo-plugins#401: This PR releases the Sandbox runtime that includes @oh-my-pi/pi-coding-agent@18.12.6.

Merge Risk: Minimal · up to bd401

The development-only dependency update does not affect production shrinkwrap metadata, and no merge-blocking issue is established.

Architecture risk: Low · up to bd401

The change affects 1 system.

Changed systems: paseo-omp

Architecture concerns
No architecture-level concerns identified.

Review details

Systems and components

• observed — paseo-omp (service) was modified; 1 changed file maps to changed impact.

Before / after behavior

• observed — Modified behavior in paseo-omp/package.json: Both @oh-my-pi/pi-agent-core and @oh-my-pi/pi-coding-agent dev dependency versions changed from 18.10.1 to 18.12.6.

Pre-merge checks: 5

Check name	Status	Explanation
Description Check		Check skipped - CodeRabbit’s high-level summary is enabled.
Title check		The title clearly identifies a dependency update to OMP SDK version 18.12.6. This matches the changeset, which upgrades the OMP agent dependencies to version 18.12.6.
Docstring Coverage		No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check. Docstring coverage is scoped to functions touched by this diff. Analyzed 0 functions across 0…
Linked Issues check		Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check		Check skipped because no linked issues were found for this pull request.

Full details: Docstring Coverage

Explanation

No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check. Docstring coverage is scoped to functions touched by this diff. Analyzed 0 functions across 0 files. (1 skipped: 1 unsupported.)

Finishing Touches: 💡 1

🛠️ Fix failing CI checks: 💡

• Commit to this branch
• Create a new PR

🧪 Generate unit tests (beta)

• Commit to this branch
• Create a new PR

• Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

Warning

Some tools did not complete. Review the errors below.

🔧 Biome (2.5.15)

paseo-omp/package.json

Biome could not lint this file: nested root configuration. Check the repository's Biome configuration and plugins.

_{Comment @coderabbitai help to get the list of available commands.}