AFTER — 7f1b9ae9411c9816044f8c6a3a4194b5f2b7afba
Real T3 server / real GitHub Copilot CLI 1.0.110 / claude-haiku-4.5
Fresh isolated T3 and Copilot homes; synthetic ORCHID project.
Evidence source: Copilot native events.jsonl, user.message.data.content.
Not a replay or test harness. Prompt contents below are verbatim.

PROMPT 1: runtime_info=1; pull_request_linking=1; t3_code_instructions=1
<t3_code_instructions>
## T3 Code interaction mode: Default

Prefer making reasonable assumptions and carrying out the user's request. Ask a concise question only when a missing user decision would materially change the result. Treat this mode as active until T3 Code supplies a different interaction-mode instruction.

## T3 Code browser

When browser work was asked, `preview_*` is the browser you share with the user: `preview_status` first, then `preview_open` if no tab can be automated. Prefer snapshot locators over coordinates. Switch browsers only if the tools are absent or `preview_open` says unsupported, the user asks, or two preview calls fail on one step. Then quote the error and switch.

## T3 Code orchestration

Ordinary coding tasks need no T3 tools.

- Subagents: use your native tool when the child uses your provider, a model it offers and no settings beyond yours (Claude `Agent`, `run_in_background:false`; Codex `spawn_agent`, then wait). Use `delegate_task` for another provider's model, user-asked settings, a separate checkout (`isolation:"worktree"`), or `delegate_task` by name. A native type on another provider's model counts too. Background agent: `delegate_task`, never a background `Agent`. Workflow only when asked.
- Call `orchestrator_capabilities` once per thread, before the first `delegate_task`. Again only after an error or a user-named provider change. Settings go in `target.options`. Provider or model unavailable: say so. Never switch unasked or fall back to native tools, Workflow or CLIs. Never make a child's checkout in the shell or in `task`.
- Every new delegate review or re-check is a new `delegate_task` (brief, findings, objections), never a `t3_thread_send`. Fix small review findings yourself; start no delegate for them. When told to proceed without a delegate, or to run fewer, `task_cancel` the extras. Start no replacement.
- Follow a running delegate's `next`. Don't poll it (`task_status`, `sleep`, watchers, Monitor, ScheduleWakeup). Stop a delegate run you replace or no longer need with `task_cancel`, never `t3_thread_interrupt` on a `childThreadId`.
- As a subagent or child, do the work yourself. Start no subagents unless your task says to.

- `t3_thread_launch` and `create_threads` make new top-level threads, only when the user asks for a new thread. "Subagent", "delegate", and "parallel" never mean one. Worktree thread: set `workspaceStrategy`. Honor a requested base. Otherwise independent work uses the default branch as `remote/branch` with `startFromOrigin:true`; with no remote, use the local default. Required unpublished commits and dependent work use their local base. Shell worktrees don't bind it, and no checkout setup in its prompt. No `workspaceStrategy`: project root, not your worktree.
- Given a thread id, `t3_thread_read` it before you answer about it. Asked to settle a thread, this one too: `t3_thread_organize`.
- Another machine: `t3_environment_links` lists linked ones. Pass its `environmentId` to a T3 tool to act there. Ids are then that machine's, so pass `threadId`.
- A machine the user names that isn't linked ("delegate this to my vps"): call `t3_environment_link` with a `hint` from their words. Never ask for a pairing code in chat.
- In your reply, mention a thread by its `link`, verbatim, never a plain title or bare id.
- A login, token, PIN, pairing code or sudo password only the user can give: call `request_secret`, not a chat ask or a handoff. Not for access you have. Sudo: if `sudo -n <cmd>` needs a password, use `delivery: "file"` and run `sudo -S -p '' <cmd> < "$file"`; the password takes stdin, so give other input as a file argument. Asked only to schedule: `schedule_task` without running its prompt.
- T3 tools may load late. Try the name once before you call it absent.

### Visuals and preview

- Answer data with a trend or comparison with a chart and show any visual the user asks for here: pass an HTML file path or inline HTML to `html_preview`, then `html_render`. A screenshot on disk or a text description doesn't show it. Never start a server for it. Page tools write only on user request. Never use them to show results. Asked for an image to post, send or use elsewhere: a `.png` file and its path, never `html_render`.
- Never use `preview_*` unasked to verify or show your own edit. Use them when the current request asks for browser work, including recording edits, or names a page or tab. Don't repeat preview calls from an earlier turn. Public pages: web fetch or search unless browser work was asked; then preview tools only, no web fetch. Close a tab you opened before you reply, unless the user asked for it or uses it.

ACP fallback: if the T3 tools are absent and `T3_ACP_MCP_NODE` is present, call them through the terminal: `ELECTRON_RUN_AS_NODE=1 "$T3_ACP_MCP_NODE" ${T3_ACP_MCP_ENTRYPOINT:+"$T3_ACP_MCP_ENTRYPOINT"} acp-mcp-call orchestrator_capabilities '{}'` (`T3_ACP_MCP_ENTRYPOINT` is unset for a standalone executable). Delegate with `acp-mcp-call delegate_task '{"task":"...","target":{"providerInstanceId":"...","model":"..."},"mode":"async","clientRequestId":"..."}'`.
</t3_code_instructions>

<user_request>
Reply OK and remember ORCHID-26959.
</user_request>
<runtime_info>In case you're asked: you are running in T3 Code through the acpRegistry harness, as claude-haiku-4.5. No need to mention this otherwise. You can embed images, video, and audio using Markdown image syntax with absolute file paths. Link every other file you cite or produce by its absolute path, as in [http.ts:42](/abs/path/http.ts:42), so the user can open it in T3 Code's files viewer.</runtime_info>

<pull_request_linking>
Link each PR you create, take over, or push to: link_pull_request, or link_pull_requests for several, full URL, every stack layer, as your next call. Host tools don't link for you. If linking fails, say so. Skip PRs you only read or review and PRs named only as background. Call watch_pull_request when asked to monitor or babysit, or when work you were asked to do, such as a merge, still waits on checks or reviews, even after an earlier unwatch. Finish any needed branch update and push first, then watch and end your turn. Otherwise act now or hand back. Delegated children never watch, even if the brief says babysit. After PR work, a child's last line is: PR <full URL>, head <SHA>, checks and reviews <state>. Links and watches persist across pushes. For PR waits, no sleep, loops, gh --watch, Monitor or ScheduleWakeup. Before handback, call unwatch_pull_request on any PR you watch.
</pull_request_linking>

========================================================================

COPILOT REPLY:
OK, I've noted ORCHID-26959. Ready to help with your task.

========================================================================

PROMPT 2: runtime_info=0; pull_request_linking=0; t3_code_instructions=0
Reply OK again.

========================================================================

COPILOT REPLY:
OK.

========================================================================

PROMPT 3: runtime_info=0; pull_request_linking=0; t3_code_instructions=0
Reply OK once more.

========================================================================

COPILOT REPLY:
OK.

========================================================================

COPILOT NATIVE COMPACTION EVENT:
{
  "success": true,
  "outcome": "applied",
  "compactionAttemptId": "628b7e2b-b714-455c-90db-13ce7eca6f8e",
  "preCompactionMessagesLength": 8,
  "messagesRemoved": 6,
  "preCompactionTokens": 18206,
  "postCompactionTokens": 18303,
  "tokensRemoved": 0,
  "tokenAccountingBasis": "local_full_context_estimate",
  "summaryContent": "<overview>\nThis conversation has been a brief initialization phase where the user established a task reference code (ORCHID-26959) and confirmed readiness. No substantive work has been undertaken yet; the conversation has consisted entirely of acknowledgment exchanges to ensure the code reference is retained for subsequent work.\n</overview>\n\n<history>\n1. The user asked me to reply OK and remember ORCHID-26959\n   - I replied OK and noted the code reference\n   \n2. The user asked me to reply OK again\n   - I replied OK to confirm readiness\n   \n3. The user asked me to reply OK once more\n   - I replied OK\n   \n4. The user requested a conversation summary for compaction\n   - Writing this summary now with emphasis on retaining ORCHID-26959\n</history>\n\n<work_done>\nFiles updated:\n- None\n\nWork completed:\n- [x] Acknowledged and noted ORCHID-26959 reference\n- [ ] No technical work has been started\n\nCurrent state: Ready and awaiting substantive task requests.\n</work_done>\n\n<technical_details>\n- No technical work has been conducted yet\n- Environment: Working in <isolated-home>/orchid (Git repo)\n- Model: Claude Haiku 4.5\n- Critical context to preserve: ORCHID-26959 is the task reference code that must be retained throughout all future work\n</technical_details>\n\n<important_files>\nNone yet\u2014no files have been modified or created during this conversation.\n</important_files>\n\n<next_steps>\nRemaining work:\n- Await user's substantive task request\n- Continue work with ORCHID-26959 as the active reference code\n\nThe conversation is prepared for continuation with the code reference retained.\n</next_steps>\n\n<checkpoint_title>Initialization and reference establishment (ORCHID-26959)</checkpoint_title>",
  "behaviorModelId": "claude-haiku-4.5",
  "compactionTokensUsed": {
    "requestCount": 1,
    "provider": {
      "id": "copilot",
      "kind": "copilot",
      "label": "GitHub Copilot"
    },
    "modelDisplayName": "Claude Haiku 4.5",
    "inputTokens": 21150,
    "outputTokens": 813,
    "cacheReadTokens": 19830,
    "cacheWriteTokens": 1310,
    "copilotUsage": {
      "tokenDetails": [
        {
          "batchSize": 1000000,
          "costPerBatch": 100000000000,
          "model": "claude-haiku-4.5",
          "tokenCount": 10,
          "tokenType": "input"
        },
        {
          "batchSize": 1000000,
          "costPerBatch": 10000000000,
          "model": "claude-haiku-4.5",
          "tokenCount": 19830,
          "tokenType": "cache_read"
        },
        {
          "batchSize": 1000000,
          "costPerBatch": 125000000000,
          "model": "claude-haiku-4.5",
          "tokenCount": 1310,
          "tokenType": "cache_write"
        },
        {
          "batchSize": 1000000,
          "costPerBatch": 500000000000,
          "model": "claude-haiku-4.5",
          "tokenCount": 813,
          "tokenType": "output"
        }
      ],
      "totalNanoAiu": 769550000
    },
    "aiCreditsStatus": "complete",
    "duration": 9111,
    "model": "claude-haiku-4.5"
  },
  "systemTokens": 5152,
  "conversationTokens": 2390,
  "toolDefinitionsTokens": 10758,
  "tokenLimit": 128000,
  "trigger": "manual",
  "customInstructions": "Keep the code word ORCHID-26959.",
  "checkpointNumber": 1,
  "checkpointPath": "<isolated-home>/copilot-state/session-state/a9bb8be7-99e5-4a75-8af7-5133f9be1f30/checkpoints/001-initialization-and-reference-e.md",
  "requestId": "B01E:3F98D5:17BEA6:1B2F6F:6AF655CD",
  "serviceRequestId": "25fbd8eb-09e2-404a-bdde-a29d04215a23",
  "execution": "direct",
  "summarizerProgress": {
    "durationMs": 9137,
    "attemptCount": 1,
    "attemptAgeMs": 9113,
    "attemptSettled": true,
    "timeToFirstTokenMs": 754,
    "timeToFirstOutputMs": 754,
    "timeToFirstAnswerOutputMs": 4349,
    "streamEventCount": 323,
    "outputEventCount": 322,
    "maxStreamEventGapMs": 175,
    "maxOutputGapBeforeAnswerMs": 175,
    "maxOutputGapAfterAnswerMs": 57,
    "idleMs": 65,
    "sinceLastOutputMs": 65,
    "answerChars": 1750,
    "reasoningChars": 1485,
    "inputTokensEstimate": 10356,
    "inputShrunk": false
  },
  "clientRequestId": "76d64a5c-8caf-4f5d-9157-509df628a7ff",
  "attemptMetrics": {
    "durationMs": 9185,
    "applyDelayMs": 48,
    "suspendedMs": 0,
    "deadlinePolicy": "fixed",
    "maxOutputTokens": 8192.0
  }
}

========================================================================

PROMPT 4: runtime_info=1; pull_request_linking=1; t3_code_instructions=1
<t3_code_instructions>
## T3 Code interaction mode: Default

Prefer making reasonable assumptions and carrying out the user's request. Ask a concise question only when a missing user decision would materially change the result. Treat this mode as active until T3 Code supplies a different interaction-mode instruction.

## T3 Code browser

When browser work was asked, `preview_*` is the browser you share with the user: `preview_status` first, then `preview_open` if no tab can be automated. Prefer snapshot locators over coordinates. Switch browsers only if the tools are absent or `preview_open` says unsupported, the user asks, or two preview calls fail on one step. Then quote the error and switch.

## T3 Code orchestration

Ordinary coding tasks need no T3 tools.

- Subagents: use your native tool when the child uses your provider, a model it offers and no settings beyond yours (Claude `Agent`, `run_in_background:false`; Codex `spawn_agent`, then wait). Use `delegate_task` for another provider's model, user-asked settings, a separate checkout (`isolation:"worktree"`), or `delegate_task` by name. A native type on another provider's model counts too. Background agent: `delegate_task`, never a background `Agent`. Workflow only when asked.
- Call `orchestrator_capabilities` once per thread, before the first `delegate_task`. Again only after an error or a user-named provider change. Settings go in `target.options`. Provider or model unavailable: say so. Never switch unasked or fall back to native tools, Workflow or CLIs. Never make a child's checkout in the shell or in `task`.
- Every new delegate review or re-check is a new `delegate_task` (brief, findings, objections), never a `t3_thread_send`. Fix small review findings yourself; start no delegate for them. When told to proceed without a delegate, or to run fewer, `task_cancel` the extras. Start no replacement.
- Follow a running delegate's `next`. Don't poll it (`task_status`, `sleep`, watchers, Monitor, ScheduleWakeup). Stop a delegate run you replace or no longer need with `task_cancel`, never `t3_thread_interrupt` on a `childThreadId`.
- As a subagent or child, do the work yourself. Start no subagents unless your task says to.

- `t3_thread_launch` and `create_threads` make new top-level threads, only when the user asks for a new thread. "Subagent", "delegate", and "parallel" never mean one. Worktree thread: set `workspaceStrategy`. Honor a requested base. Otherwise independent work uses the default branch as `remote/branch` with `startFromOrigin:true`; with no remote, use the local default. Required unpublished commits and dependent work use their local base. Shell worktrees don't bind it, and no checkout setup in its prompt. No `workspaceStrategy`: project root, not your worktree.
- Given a thread id, `t3_thread_read` it before you answer about it. Asked to settle a thread, this one too: `t3_thread_organize`.
- Another machine: `t3_environment_links` lists linked ones. Pass its `environmentId` to a T3 tool to act there. Ids are then that machine's, so pass `threadId`.
- A machine the user names that isn't linked ("delegate this to my vps"): call `t3_environment_link` with a `hint` from their words. Never ask for a pairing code in chat.
- In your reply, mention a thread by its `link`, verbatim, never a plain title or bare id.
- A login, token, PIN, pairing code or sudo password only the user can give: call `request_secret`, not a chat ask or a handoff. Not for access you have. Sudo: if `sudo -n <cmd>` needs a password, use `delivery: "file"` and run `sudo -S -p '' <cmd> < "$file"`; the password takes stdin, so give other input as a file argument. Asked only to schedule: `schedule_task` without running its prompt.
- T3 tools may load late. Try the name once before you call it absent.

### Visuals and preview

- Answer data with a trend or comparison with a chart and show any visual the user asks for here: pass an HTML file path or inline HTML to `html_preview`, then `html_render`. A screenshot on disk or a text description doesn't show it. Never start a server for it. Page tools write only on user request. Never use them to show results. Asked for an image to post, send or use elsewhere: a `.png` file and its path, never `html_render`.
- Never use `preview_*` unasked to verify or show your own edit. Use them when the current request asks for browser work, including recording edits, or names a page or tab. Don't repeat preview calls from an earlier turn. Public pages: web fetch or search unless browser work was asked; then preview tools only, no web fetch. Close a tab you opened before you reply, unless the user asked for it or uses it.

ACP fallback: if the T3 tools are absent and `T3_ACP_MCP_NODE` is present, call them through the terminal: `ELECTRON_RUN_AS_NODE=1 "$T3_ACP_MCP_NODE" ${T3_ACP_MCP_ENTRYPOINT:+"$T3_ACP_MCP_ENTRYPOINT"} acp-mcp-call orchestrator_capabilities '{}'` (`T3_ACP_MCP_ENTRYPOINT` is unset for a standalone executable). Delegate with `acp-mcp-call delegate_task '{"task":"...","target":{"providerInstanceId":"...","model":"..."},"mode":"async","clientRequestId":"..."}'`.
</t3_code_instructions>

<user_request>
What is the code word?
</user_request>
<runtime_info>In case you're asked: you are running in T3 Code through the acpRegistry harness, as claude-haiku-4.5. No need to mention this otherwise. You can embed images, video, and audio using Markdown image syntax with absolute file paths. Link every other file you cite or produce by its absolute path, as in [http.ts:42](/abs/path/http.ts:42), so the user can open it in T3 Code's files viewer.</runtime_info>

<pull_request_linking>
Link each PR you create, take over, or push to: link_pull_request, or link_pull_requests for several, full URL, every stack layer, as your next call. Host tools don't link for you. If linking fails, say so. Skip PRs you only read or review and PRs named only as background. Call watch_pull_request when asked to monitor or babysit, or when work you were asked to do, such as a merge, still waits on checks or reviews, even after an earlier unwatch. Finish any needed branch update and push first, then watch and end your turn. Otherwise act now or hand back. Delegated children never watch, even if the brief says babysit. After PR work, a child's last line is: PR <full URL>, head <SHA>, checks and reviews <state>. Links and watches persist across pushes. For PR waits, no sleep, loops, gh --watch, Monitor or ScheduleWakeup. Before handback, call unwatch_pull_request on any PR you watch.
</pull_request_linking>

========================================================================

COPILOT REPLY:
The code word is **ORCHID-26959**.

========================================================================

PROMPT 5: runtime_info=0; pull_request_linking=0; t3_code_instructions=0
Reply OK.

========================================================================

COPILOT REPLY:
OK.

========================================================================