sergiopaniego/qwen3-0.6b-pimono-gkd-lr2e5
Text Generation • 0.6B • Updated • 460
messages listlengths 2 2 |
|---|
[
{
"content": "do 1",
"role": "user"
},
{
"content": "Exposing `authStorage` on the harness and moving the text helpers over to `harness.ts`.",
"role": "assistant"
}
] |
[
{
"content": "is the word \"stuff\" in your system prompt?",
"role": "user"
},
{
"content": "No, the word \"stuff\" does not appear in my system prompt.",
"role": "assistant"
}
] |
[
{
"content": "what does package manager do concretely? where does it put the things?",
"role": "user"
},
{
"content": "Concrete PackageManager behavior proposal\n\nGlobal install locations\n- npm: use `npm root -g` to find global node_modules. If source is npm:@scope/pkg@1.2.3, install via `npm inst... |
[
{
"content": "ok, so fix good?",
"role": "user"
},
{
"content": "Functionally, yes.\n\n- tests pass\n- non-Termux height resize still does full redraw\n- Termux plain terminal sessions skip the redraw path\n- width resize behavior is unchanged\n\nBut policy-wise, it is slightly more complicated than... |
[
{
"content": "fuck again",
"role": "user"
},
{
"content": "I independently reproduced it.\n\n## Issue #2003\n**Title:** Theme hot-reloading is hitting the cache rather than reloading the file\n\n### What I read\n- Issue body and all comments\n- Linked PR: `#2214`\n- Related code, in full:\n - `pack... |
[
{
"content": "again i implore you to NOT emit control characters in your tool calls. now make the edit",
"role": "user"
},
{
"content": "Let me implement the changes. First, the sort in `package-manager.ts`:",
"role": "assistant"
}
] |
[
{
"content": "no, i told the agent to sleep, i ran the bash, which completed immediately, the agent was still sleeping, spinner spun, i resized, typed, no crash",
"role": "user"
},
{
"content": "You're right. Given that sequence, I don't have a crash repro.\n\nAt this point the honest conclusion is:... |
[
{
"content": "when do we set partialJson? if it's empty, and done args are not, we can just emit a tool_update, no? that would solve it for ths csae",
"role": "user"
},
{
"content": "Yes, if you only want to solve the reported case, the smallest patch is:\n\n- when handling `response.function_call_a... |
[
{
"content": "truncateTail(...) i would assume that's also slow, no?",
"role": "user"
},
{
"content": "Potentially, yes. If you call `truncateTail(fullText)` on every chunk, it is not free.\n\n## What it costs\n\n`truncateTail()` currently does work like:\n\n- compute total bytes of the whole string... |
[
{
"content": "ok. what if the project local settings.json has npm in it? will that also be installed to the global npm modules?",
"role": "user"
},
{
"content": "With the current decision to use npm root -g, yes. Any npm: source in project settings would still install into the global npm modules dir... |
[
{
"content": "oki, now for the chaining, propose how it would loo",
"role": "user"
},
{
"content": "Yep. I would structure it like this.\n\n### 1) Give `tool_result` its own method in runner\n\n`packages/coding-agent/src/core/extensions/runner.ts`\n\n```ts\nasync emitToolResult(event: ToolResultEven... |
[
{
"content": "@packages/tui/src/components/editor.ts has 3 lines by default, top border, input line, bottom border. when a user does auto-complete (see @packages/tui/src/autocomplete.ts) additional lines are shown below the bottom border, which adds new lines to the scrollback. when the user cancels, or selects... |
[
{
"content": "can we do a if(BUN) import(bedrock) in registration?",
"role": "user"
},
{
"content": "Not in a plain runtime `if`.\n\nExample:\n```ts\nif (typeof Bun !== \"undefined\") {\n\treturn import(\"./amazon-bedrock.js\");\n}\n```\n\nProblems:\n\n- **Bun compile**: likely good, because the lit... |
[
{
"content": "how can i press page up and down on a mac book keyboard?",
"role": "user"
},
{
"content": "On MacBook:\n\n- **Fn + Up Arrow** = Page Up\n- **Fn + Down Arrow** = Page Down\n\nAlso:\n- **Fn + Left Arrow** = Home\n- **Fn + Right Arrow** = End",
"role": "assistant"
}
] |
[
{
"content": "are you sure the missing currentBlock = null is a preexisting bug? also, what effect does this have on responses from providers that DO support tool result streaming?",
"role": "user"
},
{
"content": "Good questions. Let me re-examine.\n\n**On `currentBlock = null`:**\n\nLooking at the... |
[
{
"content": "- Bun runtime takes precedence over npm detection, so running with bun run on an npm install will suggest\n bun install -g even though npm was used.\n\nno, pi ships as a self congtained bun executable so that seems right.\n\nwhat did the submitter mean with \"looping\"?",
"role": "user"
},
... |
[
{
"content": "ok, which path to you feel more confident with?",
"role": "user"
},
{
"content": "For correctness, I feel more confident with full unified diff.\n\nOrder of confidence:\n\n1. Full unified diff\n- most reliable\n- backed by the `diff` library\n- better error behavior\n- closer to standa... |
[
{
"content": "2. TUI tracks where the focused component starts during render how?",
"role": "user"
},
{
"content": "You're right to push on this - it's the crux of the problem. Let me think through the options:\n\n### The Core Challenge\n\nTUI calls `Container.render()` which just concatenates child... |
[
{
"content": "what's the purpose of the WAYLAND_DISPLAY and DISPLAY checks?",
"role": "user"
},
{
"content": "They make the Termux special case narrower.\n\nWithout them, any environment with `TERMUX_VERSION` would skip the height-resize full redraw. That is broader than the issue report.\n\nThe iss... |
[
{
"content": "you just emit an edit tool call you idiot ...",
"role": "user"
},
{
"content": "Correct. I can call the harness `edit` tool, but I cannot emit a pi `tool_call` event into your already-running interactive session from here.\n\nThose are different paths:\n\n- my API `edit` call edits fil... |
[
{
"content": "ExtensionCommandContext is bound to by the mode, right?",
"role": "user"
},
{
"content": "Yes.\n\n`ExtensionCommandContext` is created by the extension runner, but the actual method implementations are **bound by the mode/runtime** via `bindCommandContext(...)`.\n\nFlow is:\n\n1. Mode/... |
[
{
"content": "think",
"role": "user"
},
{
"content": "I'm ready to work. Here's what I can help you with:\n\n**Code Tasks:**\n- Read, edit, and write files\n- Execute bash commands (search, git, file operations)\n- Implement features, fix bugs, refactor code\n- Run type checks and tests\n\n**Project... |
[
{
"content": "what if we add a tool result after that and do another call? that would mean we have an orphaned tool result, does that pass the openai responses api? modify the convertMessages with your fix, so the first issue passes, and modify the test to also see if an orphaned toolResult passes",
"role":... |
[
{
"content": "where are the dependencies installed to? does this override the node_modules and package-lock.json in this cwd?",
"role": "user"
},
{
"content": "Good catch. The `npm install --no-save --force` installs to the local `node_modules/` directory. It won't modify `package.json` or `package-... |
[
{
"content": "no, they are new features. for clear on shrink: this just means that the editor and footer will always stick to the bottom of the terminal, but it may lead to some redraws, which leads to some flicker. off by default",
"role": "user"
},
{
"content": "Got it. Let me add the New Features... |
[
{
"content": "https://github.com/continua-ai/tinny-pi this fork implements alterantive mode i think. can you have a look what they do and how big the changes are compared to the originalr epo here in this cwd?",
"role": "user"
},
{
"content": "I'll look at the fork and compare it to the local repo."... |
[
{
"content": "yo",
"role": "user"
},
{
"content": "Yo. What's up?",
"role": "assistant"
}
] |
[
{
"content": "page up/down now work",
"role": "user"
},
{
"content": "Excellent! The bug was that `parseKeyId()` lowercases the key identifier, so `\"pageUp\"` became `\"pageup\"`, but the switch cases were still camelCase. Fixed by changing:\n- `case \"pageUp\":` → `case \"pageup\":`\n- `case \"pag... |
[
{
"content": "ok, so this is now fixed?",
"role": "user"
},
{
"content": "Yes.\n\nFixed behavior:\n- Responses-based providers now send tool-result images in native `function_call_output.output`\n- no synthetic follow-up user message is inserted anymore\n\nAffected providers:\n- OpenAI Responses\n- ... |
[
{
"content": "so, my take aways\n- in newSession in AgentSession, we should append a thinking entry with the current thinking level?\n- not sure what to do in AgentSession.switchSession. if the session has a thinking entry, buildSessionContext would set it, no?\n- nots ure about SessionManager.buildSessionConte... |
[
{
"content": "there must be a way better solution than this. this is terrible.",
"role": "user"
},
{
"content": "You're right. The simplest solution: skip the test when clipboard isn't available (CI has no display). Let me check the current test:",
"role": "assistant"
}
] |
[
{
"content": "soo, can we install to ./pi/npm/node_modules easily? in case we find a npm: extension in the project local settings.json?",
"role": "user"
},
{
"content": "Yes. Use `npm install --prefix <cwd>/.pi/npm <pkg>` and resolve path from `<cwd>/.pi/npm/node_modules/<pkg>`. We can detect origin... |
[
{
"content": "async function runLoop(\n currentContext: AgentContext,\n newMessages: AgentMessage[],\n config: AgentLoopConfig,\n signal: AbortSignal | undefined,\n stream: EventStream<AgentEvent, AgentMessage[]>,\n streamFn?: StreamFn,\n): Promise<void> {\n let firstTurn = true;\n // Ch... |
[
{
"content": "can you briefly implement the suggested fix?",
"role": "user"
},
{
"content": "I’ll update the tool event ordering in `packages/agent/src/agent-loop.ts`, add a regression test in `packages/agent/test/agent-loop.test.ts`, and fix the agent README wording. Then I’ll run `npm run check`."... |
[
{
"content": "actually, codex, please take over",
"role": "user"
},
{
"content": "Taking over. The likely cause is that when pi started there were no extensions, so `InteractiveMode.initExtensions()` returned early and never called `session.bindExtensions()`. That means `_extensionUIContext` stayed ... |
[
{
"content": "what tool would you use to grep through the files in here?",
"role": "user"
},
{
"content": "Use the `bash` tool. I can run `rg` or `grep` from there, for example `rg \"pattern\" .`.",
"role": "assistant"
}
] |
[
{
"content": "if there are no entries in the session or due to other conditions, buildSessionContext() returns \"off\" for thinking level. that's bad. we want to pass in a default thinking level that is used instead of \"off\"",
"role": "user"
},
{
"content": "Understood. We can add optional defaul... |
[
{
"content": "same rule from streamFn (also in ai package). must return a stream, stream must emit shit as per event protocol, encoding errors in AssistantMessage",
"role": "user"
},
{
"content": "Yes.\n\nThat is the clean contract.\n\n## Agent package contract\n\nFor `packages/agent`, these should ... |
[
{
"content": "ok, how can i test this?",
"role": "user"
},
{
"content": "Create a minimal test extension:",
"role": "assistant"
}
] |
[
{
"content": "your excample sucks?",
"role": "user"
},
{
"content": "Example with concrete paths:\n\n- settings.json lives at `/home/ryota/.pi/agent/settings.json`\n- You run `pi install ~/Coding/github.com/rytswd/pi-agent-extensions/`\n- Stored entry becomes `../../Coding/github.com/rytswd/pi-agent... |
[
{
"content": "i thin 8080 is assigned to a different process? check",
"role": "user"
},
{
"content": "Checking what owns port 8080, then I’ll either kill the stale `llama-server` or move to a free port.",
"role": "assistant"
}
] |
[
{
"content": "so we can't search just by keywords?",
"role": "user"
},
{
"content": "You're right, `npm search` doesn't have a keywords-only filter. It searches across name, description, and keywords together.\n\nThere's no official npm CLI flag like `--keywords-only`. However, you can use the npm r... |
[
{
"content": "for \"foreign\" ids",
"role": "user"
},
{
"content": "Yes. For foreign IDs, that is a good rule.\n\nMeaning:\n\n- same-provider native IDs:\n - preserve as-is when they are already valid and backend-issued\n- foreign IDs:\n - do not try to transliterate the original structure\n - co... |
[
{
"content": "this might be an issue with the agent session runtime and host setup",
"role": "user"
},
{
"content": "Checking `sdk.ts` and the runtime setup path.",
"role": "assistant"
}
] |
[
{
"content": "gah",
"role": "user"
},
{
"content": "With the current logic:\n\n- `\\n\\n/` does trigger the menu because all other lines are empty.\n- `asf\\n/` does not trigger because line 0 has non whitespace.\n\nSo yes, blank lines still allow the slash menu. If you want to block `\\n\\n/` and o... |
[
{
"content": "i think we need to document the breaking change in the extension api? or is it not a breaking change?",
"role": "user"
},
{
"content": "Not a breaking change.\n\nWhy:\n- Existing extensions that return `AutocompleteItem[] | null` still work unchanged.\n- The type change is additive. It... |
[
{
"content": "ok, can we please isolate the bun garbage in coding-agent/src/bun.",
"role": "user"
},
{
"content": "Yes. That is the right place for it.\n\nClean layout:\n\n- `packages/coding-agent/src/cli.ts`\n - normal Node/npm entry\n - no Bun-specific logic\n\n- `packages/coding-agent/src/bun/c... |
[
{
"content": "oki",
"role": "user"
},
{
"content": "Not sure what you'd like to do. Pick a module (or a few) and tell me what you need, and I'll get to it.",
"role": "assistant"
}
] |
[
{
"content": "what do you think codex? easy to fix? then do it",
"role": "user"
},
{
"content": "Yes. This should be straightforward. The code already has width-aware slicing utilities, so the fix is local to `packages/tui/src/components/input.ts`.\n\nI’m going to patch that render path and then run... |
[
{
"content": "why in the fuck would you make SessionContext.thinkingLevel optional... i reverted that. also, why in the fuck did you add model to SessionContextDefaults, we agreed that we only need thinking level. i reverted that too.\n\ni think this whole approaach is fucking stupid.\n\nproblem:\n- we previous... |
[
{
"content": "ok, this too sucks. can we somehow scale the dax image down to something that doesn't hurt our download size too much, add text with a quote with daxnuts and add a link to https://mistral.ai/news/mistral-vibe-2-0 below it saying\n\nTry OpenCode\nlink",
"role": "user"
},
{
"content": "I... |
[
{
"content": "did you die?",
"role": "user"
},
{
"content": "Let me implement DL for the line deletion cases in tui.ts.",
"role": "assistant"
}
] |
[
{
"content": "why do we have to modify session manager with overrideCwd?",
"role": "user"
},
{
"content": "Because `SessionManager.open()` is the point where pi turns a session file into runtime state.\n\nRight now that method does this:\n\n- read session header\n- take `header.cwd`\n- store it as t... |
[
{
"content": "read only",
"role": "user"
},
{
"content": "I need a specific task. Let me know what to look into or which files to inspect and I’ll proceed.",
"role": "assistant"
}
] |
[
{
"content": "you absolutely can, i reloaded the extension",
"role": "user"
},
{
"content": "I still cannot call it from this session because it is not in the available tool list. If the environment exposes it as a tool, it should appear in my tool list, which it does not.",
"role": "assistant"
... |
[
{
"content": "wtf are we doing. trace if register-builtins.ts would be called before \n\n #!/usr/bin/env node\n process.title = \"pi\";\n\n await import(\"./register-bedrock.js\");\n await import(\"../cli.js\");\n\nthis was good enough i think, we do not want to duplciate any logic in ../cli.js. we can ... |
[
{
"content": "the assumption is that we only ever get a delta with EITHER content or reasoning, and keep receiving deltas of that type until the block is complete.",
"role": "user"
},
{
"content": "That assumption does not hold for some OpenAI compatible streams. Ollama can emit a delta with both co... |
[
{
"content": "is there an example extension in the pr?",
"role": "user"
},
{
"content": "No. The PR only adds the UI context methods and no example extension.",
"role": "assistant"
}
] |
[
{
"content": "i have refactored things and with it the tests. please get up to speed with the latest and greatest. can look at diffs of the last commit for example, which has all the refactoring",
"role": "user"
},
{
"content": "Checking the latest commit and changed files.",
"role": "assistant"... |
[
{
"content": "well, that binary-distribution script should do all the things necessary so we get a full distrbution like the gh workflow ...",
"role": "user"
},
{
"content": "You're right. Let me test the full script flow:",
"role": "assistant"
}
] |
[
{
"content": "you are gpt 5.4, the old analysis stems from gpt 5.1 codex mini. wanted you to double check its findings",
"role": "user"
},
{
"content": "I double-checked it independently against the code, not the issue text.\n\nI agree with the 5.1 codex mini finding.\n\n### What I verified\n\nIssue... |
[
{
"content": "ok, codex, i want you to also evaluate this. is the port faithful? do we do all the same things we used to do wrt cost tracking, message conversion, etc. pp?",
"role": "user"
},
{
"content": "Let me do a systematic, line-by-line comparison of all the key functionality:",
"role": "a... |
[
{
"content": "well, another clanker was modifiying other files",
"role": "user"
},
{
"content": "Right. Another agent had already staged files in the index, so my commit swept them in even though I only ran `git add` on my paths.\n\nCurrent state:\n- commit already created: `fdb1a9fc`\n- it includes... |
[
{
"content": "i don't understand why this worked. i start with opus, it generates a thinking thing and a tool call, then i switch to codex (openai-responses), and ... it just works? why? how?\n\n /Users/badlogic/.pi/agent/sessions/--Users-badlogic-workspaces-pi-mono--/2026-01-21T23-18-35-115Z\n _c9acbd2c-9237-4... |
[
{
"content": "so, i guess when we write to settings.json, we need to make these paths relative to the settings.json, no? or how can they be resolved otherwise if we just have the cwd relative path in settings.json, both user and project?",
"role": "user"
},
{
"content": "Right now local package path... |
[
{
"content": "i say: fuck #2391, go back to the old shit",
"role": "user"
},
{
"content": "I mostly agree.\n\nIf the choice is:\n\n1. keep global eviction and break unrelated UIs, or\n2. regress #2391 and keep behavior local and predictable\n\nI would take 2.\n\n## What I would revert\n\nNot all of ... |
[
{
"content": "is there a good point where we can do this?",
"role": "user"
},
{
"content": "Yes. There is a good internal point, but it is not a single existing extension event.\n\nThe right place is:\n\n- after `AgentSession` finishes processing an agent event\n- at the tail of its serialized `_age... |
[
{
"content": "yeah, works fine. i guess the model eventually looses track that this is NOT the tool call format, but merelly a tool result from an earlier part of the session. we should probably just adjust this message\n\nparts.push({\n text: `[Tool Call: ${block.name}]\\nArguments: ... |
[
{
"content": "do we need to rework things? host is not a thing anymore",
"role": "user"
},
{
"content": "Yes, at least partially.\n\nNot a wholesale rewrite, but some cleanup is warranted.\n\n## What still applies\nMost of the suite is still about `AgentSession` behavior and remains valid:\n\n- prom... |
[
{
"content": "yo",
"role": "user"
},
{
"content": "If you mean GPT-5.4 in context of this PR, it should be irrelevant.\n\nThis bug is capability-based, not model-name-based. The logic should be:\n\n- if current model supports reasoning, persist thinking changes normally\n- if current model does not ... |
[
{
"content": "wtf, i don't undersatnd what a kill-ring is. explain to me",
"role": "user"
},
{
"content": "A kill ring is a clipboard history for deletions. When you use delete commands like Ctrl+W (delete word), Ctrl+U (delete to line start), Ctrl+K (delete to line end), the removed text is stored ... |
[
{
"content": "fuck, i just undid the change",
"role": "user"
},
{
"content": "Understood.\n\nThe fix was in:\n- `packages/coding-agent/src/modes/print-mode.ts`\n- `packages/coding-agent/src/main.ts`\n- `packages/coding-agent/test/print-mode.test.ts`\n\nIf you want, I can reapply it now and rerun:\n-... |
[
{
"content": "what in the acutal fuck are you doing?",
"role": "user"
},
{
"content": "You're right. I'm overcomplicating this massively.\n\nThe simple fix: skip the test in CI. The clipboard requires a display, CI doesn't have one.",
"role": "assistant"
}
] |
[
{
"content": "The main design question is what happens on partial failure — if the 3rd of 5 replacements can't\n find its old_string, do you roll back the first two? Fail the whole batch? Apply what you can and\n report which failed?",
"role": "user"
},
{
"content": "Fail the whole batch. Roll ba... |
[
{
"content": "prepareForExit this is a piss poor implementation",
"role": "user"
},
{
"content": "You're right, a magic sleep is garbage. Let me think about this properly.\n\nThe real fix is to not request key release events unless something actually needs them. Currently we always enable flag 2 (re... |
[
{
"content": "100% sure we are in a good place?",
"role": "user"
},
{
"content": "Not 100%, but close.\n\n## What is solid\n\n- The reported bug is fixed in both relevant code paths:\n - `packages/coding-agent/src/core/tools/bash.ts`\n - `packages/coding-agent/src/core/bash-executor.ts`\n- There i... |
[
{
"content": "eli5 (well concise at least) wtf the epoch shit is about",
"role": "user"
},
{
"content": "Epoch = **\"which version of reality are we in?\"**\n\nUse it to ignore old async work after the world changed.\n\n### Simple example\n\nSuppose run A ends and schedules some async cleanup:\n\n``... |
[
{
"content": "why can't we use the key prefixes for that?",
"role": "user"
},
{
"content": "Because the prefixes are not actually the same thing as runtime context.\n\nThey tell you what the binding is named under, not where it is active.\n\nExamples:\n\n### Case where prefixes happen to work\nThese... |
[
{
"content": "ok, check again, what's available to you now?",
"role": "user"
},
{
"content": "\n\nLooking at my available tools right now: `Read`, `Bash`, `Edit`, `Write`, `Grep`, `find`, `ls`.\n\nSo I have the full read/write set. You toggled back to read/write mode.",
"role": "assistant"
}
] |
[
{
"content": "Resume Session (Current Folder) ◉ Current Folder | ○ All Name: All Sort: Threaded\ntab scope · re:<pattern> regex · \"phrase\" exact\nctrl+s sort · ctrl+n named · ctrl+d delete · ctrl+p path (off) · ctrl+r rename\n\n>\n\n Issue: Fork: child session is created but parent ... |
[
{
"content": "i need you to do 10-15 turns of read-only stuff",
"role": "user"
},
{
"content": "Sure, I'll explore the codebase with read-only operations. Let me start by looking at the project structure.",
"role": "assistant"
}
] |
[
{
"content": "i also told you that this won't fly, offset is line based, not byte based\n\nf (truncation.firstLineExceedsLimit) {\n // First line at offset exceeds 30KB - tell model to use bash\n const firstLineSize = formatSize(Buffer.byteLe... |
[
{
"content": "again",
"role": "user"
},
{
"content": "No, `process.env` lookups have overhead. Let me cache it:",
"role": "assistant"
}
] |
[
{
"content": "fucking idiot, fix",
"role": "user"
},
{
"content": "I’m restoring the antigravity override and leaving the Bedrock fix intact.",
"role": "assistant"
}
] |
[
{
"content": "yes, give me a full design",
"role": "user"
},
{
"content": "Yes. I would not use prefixes as the actual runtime context.\n\nI would use **capability contexts** plus a **per-component resolver**.\n\n## 1. Design goal\n\nWe need all of these to be true at once:\n\n1. User remaps should ... |
[
{
"content": "works",
"role": "user"
},
{
"content": "Good. Regression is covered in `packages/tui/test/markdown.test.ts`.",
"role": "assistant"
}
] |
[
{
"content": "got a changelog entry?",
"role": "user"
},
{
"content": "Yes, added to `packages/coding-agent/CHANGELOG.md` under `## [Unreleased]` → `### Added`:\n\n```\n- Added typed `ToolCallEvent.input` per tool with `isToolCallEventType()` type guard for narrowing built-in tool events ([#1147](ht... |
[
{
"content": "the fix is active, it's just garbage. rever tall file changes",
"role": "user"
},
{
"content": "Reverting all changes from this session, including the agent fix and the repro extension files.",
"role": "assistant"
}
] |
[
{
"content": "dude, think LONG",
"role": "user"
},
{
"content": "Long version:\n\nThis repo is a full-stack agent platform, not just a single CLI tool. \nIts purpose is to make AI agent development composable across runtime, model access, interfaces, and operations.\n\nCore intent by layer:\n\n- `p... |
[
{
"content": "don't think that's the model id, it's gpt-5.4, try it",
"role": "user"
},
{
"content": "\n\nLet me check if that file exists already, then create or update it.",
"role": "assistant"
}
] |
[
{
"content": "ok, so i agree that updates of unpinned packages should follow what we do for git, that is require a pi update call. but i also think that we should do a background check for each package (npm and git) and see if there are updates. and if so, show a thing in interactive mode like we do for pi upda... |
[
{
"content": "\"this looks like a github repo path (user/repo), I should git clone it\"\n\nwhy in the fuck would pi do that??",
"role": "user"
},
{
"content": "Because `pi install` accepts git shorthands, and the parser is too permissive.\n\nIt intentionally supports inputs like:\n\n- `github.com/us... |
[
{
"content": "ok, delete llama cpp and the model cache shit",
"role": "user"
},
{
"content": "Removing:\n- `~/src/llama.cpp`\n- Hugging Face cache for `ggml-org/gemma-4-26b-a4b-it-GGUF`\n\nI’ll stop any running `llama-server` first.",
"role": "assistant"
}
] |
[
{
"content": "don't care. gpt-5.2 is all we need to check for",
"role": "user"
},
{
"content": "If you only care about `gpt-5.2`, the simplest is to change `supportsXhigh` to return true when `model.id.includes(\"gpt-5.2\")` or equals any existing set entries. That would live in `packages/ai/src/mod... |
[
{
"content": "ello there",
"role": "user"
},
{
"content": "Hey! Let me get oriented.",
"role": "assistant"
}
] |
[
{
"content": "Wrap it.\n\nAdditional instructions: \n\nDetermine context from the conversation history first.\n\nRules for context detection:\n- If the conversation already mentions a GitHub issue or PR, use that existing context.\n- If the work came from `/is` or `/pr`, assume the issue or PR context is alread... |
[
{
"content": "don't care",
"role": "user"
},
{
"content": "I need a pattern and a path to run rg. Provide both.",
"role": "assistant"
}
] |
[
{
"content": "i want to sponsor some of our dependencies in pi-mono. can you check which projects we depend on that ask for sponsorship and aren't huge?",
"role": "user"
},
{
"content": "I'll check the dependencies across the monorepo and look for sponsorship information. Let me start by examining t... |
[
{
"content": "give me a link for taht file on github",
"role": "user"
},
{
"content": "https://github.com/openai/codex/blob/main/codex-rs/core/gpt_5_2_prompt.md",
"role": "assistant"
}
] |
[
{
"content": "running via the vitest extension in vs code",
"role": "user"
},
{
"content": "The VS Code Vitest extension can have stale state or caching issues. Try:\n\n1. **Reload VS Code window** (Cmd+Shift+P → \"Developer: Reload Window\")\n2. Or **restart the Vitest extension** (click the Vitest... |
End of preview. Expand in Data Studio
README.md exists but content is empty.
- Downloads last month
- 81