## Why
Sandboxed process launch through the exec server was unsupported on Windows.
## What changed
- Add a shared native process launcher that selects the Windows sandbox session backend when required and otherwise launches PTY or pipe-based processes.
- Pass Windows permission, workspace-root, proxy, filesystem-override, and desktop settings through exec-server process preparation.
- Preserve piped stdin and inherited file descriptors across the shared launch path.
## Testing
- Verify that a remote Windows sandbox process accepts stdin while blocking a write under a read-only policy.
- Enable workspace-root sandbox tests on native Windows while continuing to skip environments that cannot emulate Windows sandbox semantics.
GitOrigin-RevId: 6719ccbe87ad0d5a57af56503e3f23f894c60a71
## What changed
- Remove the `spawn_agents_on_csv` and `report_agent_job_result` tools and their agent-job runtime and state models.
- Drop the legacy `agent_jobs` and `agent_job_items` tables during state database migration.
- Keep `features.enable_fanout` and `agents.job_max_runtime_seconds` accepted as no-op compatibility settings while omitting them from the generated configuration schema.
## Testing
- Verify upgrades remove both legacy agent-job tables.
- Verify the removed feature and configuration keys still parse without taking effect.
GitOrigin-RevId: 8cc3337da78c67162229f02f40a747f503542646
## What changed
- Make the MCP connection manager's event sender optional.
- Skip startup notifications when a caller has no session event stream.
- Decline interactive MCP elicitations in that mode while allowing the
underlying operation to continue.
- Update threadless resource reads, status snapshots, and connector discovery
to use the channel-free path.
## Testing
- Extend the app-server MCP resource test to verify that a threadless resource
read declines an elicitation and still returns the resource contents.
GitOrigin-RevId: 65fea152adbc02a78e09a40cae4134b6dccffd34
## What changed
- Assert truncated audio history through `ContextManager::raw_items()`.
- Initialize `additional_context_limit` in the quoted Windows hook command test.
GitOrigin-RevId: 43c19804e558b7a11fc17590506ec6ec7f3a9897
## What changed
- Change `ReviewDecision::Denied` to carry a rejection string and update the generated protocol schemas.
- Preserve specific rejection reasons through command, patch, network, MCP, delegated, and automatic approval flows so tool results can return them to the model.
- Distinguish invalid approval responses from user declines and truncate oversized rejection messages before adding them to model context.
## Testing
- Add coverage for denial serialization, invalid app-server approval responses, dropped automatic reviews, network rejection propagation, and rejection-message truncation.
GitOrigin-RevId: d48a55643aa03c7ecfa08a8e4b669913eec5464a
## Why
Mid-turn auto-compaction queued `SessionStart` hooks but continued sampling
before running them. This delayed hook-provided context and ignored requests to
stop the continuation.
## What changed
Drain pending session-start hooks immediately after successful mid-turn
auto-compaction. End the turn when a hook requests a stop; otherwise include
its additional context in the next sampling request.
## Testing
Added coverage for repeated compactions in one turn, context delivery without
leaking hooks into the next user turn, and stop requests that block sampling.
GitOrigin-RevId: c57708a792fb47d98d95c38d7d91bcd9f235be84
## What changed
- Add `additionalContextLimit` to command hook configuration for events that can emit `additionalContext`.
- Apply the limit independently to each hook's context before it is sent to the model. Unset values retain the 2,500-token default, while `0` disables spilling for that hook.
- Expose configured limits through app-server hook and requirements responses, and show them in the TUI hooks browser.
## Testing
- Cover JSON and TOML parsing, discovery and hashing, per-hook spilling behavior, app-server responses, and TUI rendering.
GitOrigin-RevId: 55ddc2dbbdf9067dceee3e1968c03617bbf6bcea
## Why
Cloning a `ContextManager` deep-copies every `ResponseItem`, even when callers only need a read-only history snapshot.
## What changed
- Store history items in an `Arc<Vec<ResponseItem>>` so cloned snapshots share their items until one is mutated.
- Reuse an owned vector when consuming an unshared snapshot and clone it only when other snapshots still reference it.
- Estimate session token counts from a snapshot after releasing the session-state lock.
## Testing
Added a regression test that verifies cloned histories share their backing items before mutation and diverge without affecting each other afterward.
GitOrigin-RevId: 24605888209ae9ff78cadb04d9c36463f7e2dbe0
## What changed
- Point the default Codex Apps MCP server at `ps/mcp` instead of the legacy Apps endpoint for both `backend-api` and `api/codex` base URLs.
- Use the same server configuration for Codex Apps and the hosted plugin runtime.
- Update MCP integration tests and test-server routing to expect the plugin-service path.
GitOrigin-RevId: 939f20dcff67ba6f79c11b328bfc624b25e3aac2
## Why
Audio attachments could be dropped from tool outputs and reconstructed thread history, while inline base64 payload size could overstate their model context cost.
## What changed
- Preserve remote and local audio attachments in legacy user-message events and thread history.
- Keep audio from function and custom tool outputs when the model supports it and the output fits the truncation budget; otherwise replace it with an omission marker.
- Estimate inline audio cost from its duration, with a serialized-size fallback for malformed or unsupported data URLs, so compaction and truncation use model-visible cost instead of raw base64 size.
- Replace audio in messages and tool outputs with an explanatory placeholder for models without audio input support.
## Testing
Add coverage for history round trips, dynamic tool responses, modality filtering, duration-based context estimates, output-budget truncation, and remote compaction.
GitOrigin-RevId: 5ed4628850a668f8383b7cd73db71b144cc02cd1
## Why
Building a Responses WebSocket request copied the model, instructions, input
history, tools, and other request settings into a second payload before
serialization.
## What changed
- Make `ResponseCreateWsRequest` borrow shared fields from
`ResponsesApiRequest`.
- Keep only WebSocket-specific fields owned, and allocate a separate input
vector only for incremental continuations.
- Restore response item IDs after preparing a full request so the request
retained for continuation matching keeps its original state.
## Testing
Update the serialization test to verify that the borrowed WebSocket request
produces the same wire payload as the Responses API request plus its
WebSocket-specific fields.
GitOrigin-RevId: b63d746b7f5164b5204ab95b057c914a4aeb71e1
## What changed
Remove the fallback that replaced image content in the latest tool output with
`Invalid image` and retried the model request. Invalid-image responses now use
the existing bad-request error path without modifying conversation history.
GitOrigin-RevId: f4ea782f16c5451feb261694caa73b77f24c3e4e
## What changed
- Tell the model that generated images are already displayed to the user and should not be repeated as Markdown images or file links.
- Keep generated-image artifact paths and output hints owned by the image generation extension.
- Centralize the availability check for the image generation tool across feature, provider, namespace-tool, model modality, and authentication requirements.
## Testing
- Verify artifact paths sanitize session and call identifiers.
- Verify the model-facing output hint reports that the generated image is already displayed.
GitOrigin-RevId: 6c865b6a96d6a71bb2c33246d1518a00e7c8b840
## What changed
- Remove unused APIs and test helpers across the Rust workspace.
- Route Linux sandbox test setup through the shared exec request path.
- Drop dependencies that became unnecessary after the cleanup.
GitOrigin-RevId: 48fd9b2b1575bad7add0819786b4d8ec88d8486f
## Why
zsh prints tied scalar/array exports such as `PATH` and `path` with `-T`
metadata. The snapshot filter rejected that syntax, so commands available only
through the configured `PATH` could not be resolved from a restored snapshot.
## What changed
- Recognize writable tied exports and retain them in zsh shell snapshots.
- Skip readonly exports so restoring a snapshot does not make variables such as
`PATH` immutable.
## Testing
Added macOS coverage for restoring a tied `PATH`, handling entries with spaces,
deduplicating path entries, excluding readonly tied exports, and resolving a
command available only through the restored snapshot.
GitOrigin-RevId: 297cfae3368230c7f6fe5ef84f0a10b35a2c802c
## What changed
- On session startup, remove exact `allow` entries from `rules/default.rules`
for command prefixes that Codex no longer suggests as policy amendments.
- Record the migration in `.sandbox_migration` so it runs only once, preserving
rules created after the migration.
- Skip the migration when user and project exec policy rules are ignored.
- Expand the protected prefix list across shells, interpreters, package runners,
and destructive or privilege-related commands.
## Testing
- Cover selective removal, case-insensitive matching, one-time behavior, and the
startup path with ignored policy rules.
GitOrigin-RevId: a0c60e3f82b9630e621fd034b40462e3ab775102
## What changed
- Add an optional `initialItems` field to `thread/realtime/start` for seeding
realtime V3 session history with `user`, `developer`, and `assistant` text.
- Encode the items in Frameless Bidi session bootstrap payloads for websocket
and WebRTC transports while preserving existing payloads when the list is
omitted or empty.
- Reject non-empty initial items for V1 and V2, more than 128 items, any item
over 8,192 estimated tokens, or a combined total over 8,192 estimated tokens.
## Testing
- Cover protocol serialization, app-server forwarding, websocket and WebRTC
payloads, empty-list compatibility, version checks, and size limits.
GitOrigin-RevId: 09af367f9a07543e3f8d59b7fa5387f932dc5a6a
## What changed
- Add `audio` to the model input modalities exposed by the protocol and generated app-server schemas.
- Preserve audio in prompts for models that advertise audio input, and replace historical audio with an omission marker for models that do not.
## Testing
- Cover prompt normalization with and without audio support.
- Extend the model-switching integration test to verify that switching from a multimodal model to a text-only model strips prior image and audio content.
GitOrigin-RevId: c18e47ad602720216303ff9abb5bca4d508391ee
## What changed
- Add structured `turn_id`, `retries`, `max_retries`, and `sampling_error` fields to sampling-stream retry warnings.
- Add a focused test that captures the warning and verifies its retry and error context.
GitOrigin-RevId: dea868875d162b3f97b3c8a256359278f7267c1c
## What changed
- Add `tui.resume_cwd` with `current` and `session` modes for resume and fork flows.
- Let users persist either mode directly from the working-directory prompt, while preserving one-time choices when no preference is configured.
- Apply the preference consistently during startup and in-app resume, honor explicit `--cd` overrides, and reject unsupported or unavailable directory choices with a clear error.
## Testing
- Cover configuration parsing and persistence, prompt behavior, startup and in-app flows, session metadata fallback, and remote workspace constraints.
GitOrigin-RevId: 7bf144c4b8c55bdb868cc8a71caf397c7740451a
## What changed
- Model permission instructions as a world-state section keyed by a stable hash of their rendered developer message.
- Re-emit permission context when its model-visible contents change or the retained fragment is missing, while avoiding duplicates when matching instructions already exist in history, including bundled developer messages.
- Normalize CRLF line endings when hashing world-state fragments so equivalent content produces the same snapshot.
## Testing
- Add snapshot coverage for unchanged, changed, absent, and unknown permission state.
- Cover bundled retained permissions and restored world-state baselines during rollout reconstruction.
GitOrigin-RevId: 7d8d212460c1690ca617859f66dcd825987f1e23
## Why
Audio variants were available in the user-input protocol but were replaced with unsupported-input placeholders instead of being sent to the model.
## What changed
- Serialize audio data URLs as `input_audio` content and convert local `wav`, `mp3`, `m4a`, `webm`, and `ogg` files to labeled data URLs.
- Validate and canonicalize base64 audio before requests, enforce the 50 MiB input limit, and replace invalid, unsupported, or oversized audio with explanatory text.
- Preserve audio attachments when mapping response items back to user-message events and document the app-server input variants.
## Testing
- Add unit coverage for local-file conversion, data URL validation, size and format errors, event mapping, and attachment extraction.
- Add client tests that verify data URL and local audio payloads sent to the Responses API.
GitOrigin-RevId: f72cd6b6e43ab12757eeb47621b1d670594ac7d9
## What changed
- Add `HistoryPosition` to identify a source thread and an exclusive rollout
prefix by ordinal and JSONL byte offset.
- Add optional `history_base` metadata to `SessionMeta` so a thread can record
an inherited prefix of another paginated rollout.
- Default missing `history_base` metadata to `None` and omit it when unset to
preserve compatibility with existing rollout metadata.
## Testing
- Verify that legacy session metadata without `history_base` deserializes with
no inherited history position.
GitOrigin-RevId: 3b2b07a2dfee793d3b37b0a60529f856f167dbc1
## What changed
- Add remote and local audio variants to core and app-server user inputs, including generated JSON and TypeScript schemas.
- Preserve audio fields when converting app-server inputs and response content into thread items.
- Keep unsupported audio out of model requests, TUI rendering, text extraction, and truncated tool output, using explicit text placeholders where audio is submitted as a model input.
## Testing
- Cover app-server media conversion, unsupported-input placeholders, TUI omission, and tool-output truncation.
GitOrigin-RevId: 16152bf3af674575ce569764a8b472e032e7569d
## Why
Remote executions need managed-network proxy listeners in the executor so their
loopback proxy addresses are reachable by the launched process.
## What changed
- Add a capability-gated exec-server protocol field for executor-local proxy
launch configuration, including network policy, audit metadata, and execution
attribution.
- Start the proxy while preparing a remote process, replace inherited proxy
environment variables with its local addresses, and derive the sandbox
context from its listeners.
- Keep the proxy alive until inherited output streams close, then shut it down.
- Reject unsupported remote settings such as MITM and credential injection.
## Testing
- Cover configuration round trips and rejection of unsupported settings.
- Verify executor-local startup, blocked-domain enforcement, protocol
compatibility, and proxy lifetime through process closure.
GitOrigin-RevId: c984f54e3e600aa9ebcbf8cf4574046e2c199d11
## What changed
- Add `codexResponseHandoffMode` to `thread/realtime/start` for V3 sessions, with `thinking` as the default and `commentary` and `bemTags` routing modes.
- Map BEM `analysis` and `commentary` output to the commentary channel and `final` output to the speakable channel, while preserving the original BEM envelope. Unrecognized output falls back to speakable.
- Mark explicit speech appends as speakable and stop adding the legacy `"Agent Final Message"` prefix to V3 handoffs. V1 and V2 continue to ignore the new setting.
## Testing
- Cover channel selection for every routing mode, streamed BEM header parsing, fallback behavior, explicit speech, and V1 compatibility.
GitOrigin-RevId: 1974578257488f2914b252c9c1990ba38831c96b
## What changed
- Add the `SessionEnd` hook event to hook configuration, protocol schemas, analytics, and the hooks browser.
- Run matching hooks for root threads during shutdown, including app-server archive, delete, idle unload, and graceful shutdown. Flush the transcript first and provide the session ID, transcript path, working directory, and `reason: "other"`.
- Keep teardown bounded: hook output is advisory, the default timeout is one second, configured timeouts are capped at three seconds, and async hooks run synchronously with a warning.
## Testing
- Cover transcript availability, matcher selection, timeout normalization, ignored control output, subagent exclusion, archive/delete ordering, and graceful shutdown of multiple loaded threads.
GitOrigin-RevId: 5d19a658677a137caf836ed5042dcb43f5eb6d1a
## What changed
- Represent whether realtime is active as a persisted world-state section.
- Emit start and end instructions from world-state transitions, including when
reconstructing context after resume.
- Retain existing realtime start fragments and avoid emitting another start when
custom instructions change while realtime remains active.
## Testing
- Add snapshot coverage for absent, unknown, active, and inactive realtime state
transitions.
- Cover realtime start and end updates plus resuming with changed custom start
instructions.
GitOrigin-RevId: ec670b8132b0f1c9317738bfdeb272069927d91d
## What changed
- Add a thread-owned `McpRuntime` that atomically publishes connection-set
replacements and owns shutdown.
- Route `McpResourceClient` snapshots and cache keys through that runtime so
extension clients observe refreshed MCP servers.
- Keep model-step configuration snapshots separate from ownership of live MCP
connections.
## Testing
- Add coverage that captures an extension's session resource client, refreshes
the MCP server configuration, and verifies that the existing client sees the
newly published server.
GitOrigin-RevId: 351b2e25594b8eb6b91e50ff8d06b7d49768ae1c
## What changed
- Model collaboration mode as a world-state section whose persisted snapshot is
the active mode.
- Emit collaboration instructions when the mode changes, while ignoring
instruction text changes within the same mode.
- Restore persisted collaboration instructions when they are missing from
retained history, including after a fork.
## Testing
- Add snapshot coverage for initial, unchanged, changed, and unknown mode state.
- Cover retained-history restoration and same-mode instruction changes.
GitOrigin-RevId: ffd1683499f09acef6516b0bfcb5bcb4d3f28df4
## What changed
- Remove the ignored explorer-role test for role-specific model and reasoning defaults.
- Remove two ignored unified exec tests for timeout capping and completed-session cleanup.
GitOrigin-RevId: 39386546c7628c2e1cf713139cd09fd0bf172700
## What changed
- Remove `ToolRegistry::dispatch_any`, which always forwarded a missing terminal outcome.
- Update registry and dispatch trace tests to call `dispatch_any_with_terminal_outcome` with `None` explicitly.
GitOrigin-RevId: aa3acb2ed480fd452bc25301a351f46e759bbfab
## What changed
- Treat an empty custom `multi_agent_mode_hint_text` as having no instructions to emit.
- Avoid adding an empty multi-agent mode message both on the initial turn and when the configured hint changes after resuming a session.
## Testing
- Update the empty-hint integration test to assert that no mode message is sent.
- Add resume coverage that verifies changing a custom hint to an empty string emits no update.
GitOrigin-RevId: 8033292115936fac2cf22f85141278595804ee61
## What changed
Add a Unix integration test that starts two `exec-server` environments with
separate workspace roots. Verify that each environment can write within its own
workspace while cross-workspace writes are denied.
GitOrigin-RevId: 98f2f58b97c1dcca8e103702b067c65d4eeed0c4
## Why
The shared core test harness does not need to spawn a login shell for every
test, and approval scenarios exercise tool behavior rather than rollout
persistence.
## What changed
- Disable `ShellSnapshot` by default in `test_codex()`; snapshot-specific tests
can continue to opt in explicitly.
- Give each approval scenario a named in-memory thread store so it does not use
durable rollout storage.
GitOrigin-RevId: 787670857aa4f77d5fc442ec1e0bfe6d22314f63
## What changed
- Stream V3 agent message deltas to `delegation.context.append` at 200 ms intervals instead of waiting for each completed message.
- Send commentary without a prefix and label final or phase-less V1/V3 output with `"Agent Final Message":`.
- Bound streamed output while preserving both its beginning and final tail when truncation is required.
- Remove `codexResponseHandoffPrefix` from `thread/realtime/start`; unknown fields from older clients remain ignored.
## Testing
- Cover incremental V3 commentary and final output, Unicode-safe truncation, V1/V3 phase labeling, and legacy request compatibility.
GitOrigin-RevId: 633c18b5eb9a4f5756a264b63b7bfc185b073951
## What changed
- Add `<source>transcript_tail_flush</source>` to realtime delegation payloads
emitted when a session flushes its remaining transcript tail.
- Keep explicit realtime handoff payloads unchanged.
- Render both delegation variants through a shared contextual user fragment.
## Testing
- Update realtime conversation tests to cover tagged transcript-tail flushes and
preserve the existing handoff format and XML escaping behavior.
GitOrigin-RevId: 33f70d16cf45b3af8068565c4cf2b1785be3f1d9
## Why
Selected capability roots can contribute plugins, MCP servers, connectors, and
skills. Discovering each contribution separately requires repeated access to the
executor filesystem.
## What changed
- Add the `capabilityRoots/discoverV1` exec-server RPC to scan selected roots and
materialize recognized plugin manifests, configuration files, skill
instructions, and skill metadata in one bounded request.
- Add the opt-in `executor_capability_discovery` feature, with a thread-scoped
cache and per-step snapshot shared by MCP and skill discovery.
- Parse MCP, connector, and skill contributions from the materialized snapshot,
including serving cached skill instructions without another filesystem read.
## Testing
- Cover discovery limits, manifest precedence, root-local failures, cache reuse,
plugin contributions, and parity with the existing environment skill loader.
GitOrigin-RevId: f98fd2321cafb58c596db02da1f83c09d8eb375d
## What changed
- Allow the built-in `amazon-bedrock` provider to override `base_url`, `auth`,
and `http_headers` in addition to its AWS profile and region.
- Use command-based bearer authentication and configured endpoints without
applying AWS request signing, while retaining regional endpoint resolution
for the default Bedrock configuration.
- Replace the Bedrock account `credentialSource` enum with the
`usesCodexManagedCredentials` boolean so command-authenticated and other
externally managed configurations are reported consistently.
## Testing
- Cover configuration merging and validation, command-authenticated proxy
requests with custom headers, and account reporting for managed and external
credentials.
GitOrigin-RevId: d1acbe602060470583b5e12f8d304bee5be46f4c
## Why
Standalone web search and image requests need to preserve the trusted,
thread-scoped originator used for billing attribution.
## What changed
- Seed extension data with the resolved thread originator and recognize
`codex_work_cca` and `chatgpt_cca` service names as originators.
- Forward non-default originators from the web search and image generation
extensions with the shared originator-header helper.
## Testing
- Cover originator resolution, header validation, and forwarding for web
search, image generation, and image editing.
GitOrigin-RevId: 165dace8e4d24e4bd92f234bff5ec77a21a7ff6e
## Why
Changes to the selected working directory can select different `AGENTS.md`
instructions. These updates need to reach the model even when the deferred
executor feature is disabled.
## What changed
- Refresh environment readiness and `AGENTS.md` before capturing each step.
- Record model-visible world-state changes for every session instead of gating
them on the deferred executor feature.
- Preserve the initial environment context without duplicating it.
## Testing
- Cover `AGENTS.md` replacement after a working-directory change.
- Verify that initial environment context remains unique with the deferred
executor both enabled and disabled.
GitOrigin-RevId: ea5dbe590cc049170259e7eccb6a3346fb2ed8e6
## What changed
- Accept image output from `image()` and `generatedImage()` only when its URL
uses the `data:` scheme.
- Preserve the dedicated error for remote HTTP URLs and report other malformed
or unsupported image URLs as invalid image output.
## Testing
- Add service-level coverage for rejecting invalid output from both image
helpers.
- Add an end-to-end code-mode test that verifies the tool call fails instead of
returning an invalid image item.
GitOrigin-RevId: beaf8c8830574150e8166b6ff5daf7f6dc4dc0a1
## Why
With deferred execution enabled, rebuilding a step context from the latest
thread environment snapshot could apply a settings update to a turn that was
already in progress.
## What changed
- Preserve each turn's captured environment selections when creating later
step contexts.
- Refresh only the readiness of environments that were already starting, while
retaining their original selection order.
- Apply updated environment selections to the next turn.
## Testing
Add coverage for changing the selected workspace while a turn is paused,
verifying that the active turn keeps its original working directory and the
next turn uses the new one.
GitOrigin-RevId: c6f4388c843384ef045b7abdbc303dcfca11d230
## Why
Durable v2 sub-agents can be lazily reloaded after their root session resumes. The reload path restored the agent identity but did not reapply the agent's selected role configuration.
## What changed
- Recover the agent role from the resumed session source and apply it before reloading the sub-agent.
- Preserve the runtime approval policy, approval reviewer, working directory, and permission profile while applying the role.
## Testing
Extend the cold-resume integration test to verify that the reloaded worker retains its role-defined instructions, model, provider, reasoning effort, and permission profile.
GitOrigin-RevId: a26dbbd2599838ca317326a1223a6a0a9cb2a2ed
## Why
Agent roles can override the model or reasoning effort after the initial
`spawn_agent` configuration is validated, leaving an unsupported combination
unchecked.
## What changed
- Validate the final reasoning effort against model metadata whenever a role
changes the spawned agent's model settings.
- Preserve independently configured subagent defaults when a role does not
override them.
- Apply the same role validation in both multi-agent spawn implementations.
## Testing
Added coverage for independent model and reasoning defaults, instruction-only
roles, and rejection of a reasoning effort unsupported by a role-selected
model.
GitOrigin-RevId: 1885314e9380f778603f067fc04a6225ee963428
## What changed
- Allow parallel `write_stdin` tool calls to interact with independent terminal sessions.
- Serialize reads and writes within each session so its output buffer and process lifecycle remain consistent.
- Prevent active sessions from being pruned while a `write_stdin` interaction is in progress.
## Testing
- Added unified exec and code mode coverage that uses two terminal sessions concurrently and verifies output from both.
GitOrigin-RevId: 618f5503f4c4670421f680c0f0c27414bc6b2ddd
## Why
An execution environment marked `starting` is not yet usable, but unrelated work can continue while it becomes ready.
## What changed
- Add developer guidance that explains execution-environment isolation and the resources that become available after startup.
- Tell the model to wait only when the current task needs the starting environment.
- Retain the guidance across world-state updates without duplicating it in request history.
## Testing
- Cover world-state transitions, history retention, rollback handling, and deferred-environment request deduplication.
GitOrigin-RevId: f12f9204d4cbddce50a70eb0998beb6b66cd2fa9
## What changed
- Tell the model to call `wait_for_environment` only when a task needs the selected environment's files, commands, or installed capabilities.
- Document that waiting can take several minutes and blocks other tool calls, and direct the model to continue if startup fails.
- Require the exact `starting` environment ID shown in `<environment_context>`.
GitOrigin-RevId: 6b434b835d379318879d5499c365c70598673afd
