Go's html/template has a built-in security filter (urlFilter) that only
allows http, https, and mailto URL schemes. Any other scheme gets
replaced with #ZgotmplZ.
The OAuth2 app's callback URL uses custom URI scheme which the filter
considers unsafe. For example the Coder JetBrains plugin exposes a
callback URI with the scheme jetbrains:// - which was effectively
changed by the template engine into #ZgotmplZ. Of course this is not an
actual callback. When users clicked the cancel button nothing happened.
The fix was simple - we now wrap the apps registered callback URI into
htmltemplate.URL. Usually this needs some validation otherwise the
linter will complain about it. The callback URI used by the Cancel logic
is actually validated by our backend when the client app
programmatically registered via the dynamic OAuth2 registration
endpoints, so we refactored the validation around that code and re-used
some of it in the Cancel handling to make sure we don't allow URIs like
`javascript` and `data`, even though in theory these URIs were already
validated.
In addition, while testing this PR with
https://github.com/coder/coder-jetbrains-toolbox/pull/209 I discovered
that we are also not compliant with
https://www.rfc-editor.org/rfc/rfc6749#section-4.1.2.1 which requires
the server to attach the local state if it was provided by the client in
the original request. Also it is optional but generally a good practice
to include `error_description` in the error responses. In fact we follow
this pattern for the other types of error responses. So this is not a
one off.
- resolves#20323
<img width="1485" height="771" alt="Cancel_page_with_invalid_uri"
src="https://github.com/user-attachments/assets/5539d234-9ce3-4dda-b421-d023fc9aa99e"
/>
<img width="486" height="746" alt="Coder Toolbox handling the Cancel
button"
src="https://github.com/user-attachments/assets/acab71a6-d29c-4fa9-80ba-3c0095bbdc8f"
/>
<!--
If you have used AI to produce some or all of this PR, please ensure you
have read our [AI Contribution
guidelines](https://coder.com/docs/about/contributing/AI_CONTRIBUTING)
before submitting.
-->
\`InlineMarkdown\` and \`MemoizedInlineMarkdown\` lived in
\`Markdown.tsx\`
alongside a static \`import { Prism as SyntaxHighlighter } from
"react-syntax-highlighter"\` — the full PrismJS build with ~300 language
grammars. Because \`DashboardLayout\` eagerly imports
\`AnnouncementBannerView → InlineMarkdown\`, every authenticated page
loaded and evaluated the entire Prism/refractor bundle on startup even
though syntax highlighting is only used in secondary views.
This PR moves \`InlineMarkdown\` and \`MemoizedInlineMarkdown\` into
their
own \`InlineMarkdown.tsx\` file that depends only on \`react-markdown\`
and
updates all six consumers to import from the new module.
\`Markdown.tsx\`
keeps the PrismJS import for the full \`Markdown\` component, which is
only reached through lazy-loaded routes.
> 🤖 Generated by Coder Agents
Reorder error checks in isRetryableError so IsConnectionError is evaluated before context.DeadlineExceeded. Dial timeouts (*net.OpError wrapping DeadlineExceeded) were incorrectly treated as non-retryable, causing Coder Connect to fail immediately on broken tunnels with valid DNS despite existing retry logic.
Fixes#24201
Add the five REST endpoints for managing user secrets, SDK client
methods, and handler tests.
Endpoints:
- `POST /api/v2/users/{user}/secrets`
- `GET /api/v2/users/{user}/secrets`
- `GET /api/v2/users/{user}/secrets/{name}`
- `PATCH /api/v2/users/{user}/secrets/{name}`
- `DELETE /api/v2/users/{user}/secrets/{name}`
Routes are registered under the existing `/{user}` group with
`ExtractUserParam`. The delete query was changed from `:exec` to
`:execrows` so the handler can distinguish "not found" from success
(DELETE with `:exec` silently returns nil for zero affected rows).
## What
Bumps `coder/tailscale` to
[`e956a95`](https://github.com/coder/tailscale/commit/e956a950740bd737c55451f56e77038f7430a919)
([PR #113](https://github.com/coder/tailscale/pull/113)) to pick up the
`RTM_MISS` fix for the Darwin network monitor.
Already released on `release/2.31` as v2.31.8. (#24185) to unblock a
customer. This PR is to update `main`.
## Why
On Darwin, `RTM_MISS` route-socket messages (fired on every failed route
lookup) were not filtered by `netmon`, causing each one to be treated as
a `LinkChange`. When netcheck sends STUN probes to an IPv6 address with
no route, this creates a self-sustaining feedback loop: `RTM_MISS` →
`LinkChange` → `ReSTUN` → netcheck → v6 STUN probe → `RTM_MISS` → …
The loop drives DERP home-region flapping at ~70× baseline, which at
fleet scale saturates PostgreSQL's `NOTIFY` lock and causes coordinator
health-check timeouts.
The upstream fix adds a single `if msg.Type == unix.RTM_MISS { return
true }` check to `skipRouteMessage`. This is safe because `RTM_MISS` is
a lookup-path signal, not a table-mutation signal — route withdrawals
always emit `RTM_DELETE` before any subsequent lookup can miss.
Of note is that this issue has only been reported recently, since users
updated to macOS 26.4.
Relates to ENG-2394
## Summary
Exposes `credential_kind` and `credential_hint` on AI Bridge session
threads, making credential metadata visible in the session detail API.
Each thread in the `/api/v2/aibridge/sessions/{session_id}` response now
includes:
- `credential_kind`: `centralized` or `byok`
- `credential_hint`: masked credential (e.g. `sk-a...pgAA`)
Values are taken from the thread's root interception.
## Changes
- `codersdk/aibridge.go`: Added `CredentialKind` and `CredentialHint`
fields to `AIBridgeThread`
- `coderd/database/db2sdk/db2sdk.go`: Populated from root interception
in `buildAIBridgeThread`
- `SessionTimeline.stories.tsx`: Added fields to mock thread data
The startup-timeout integration tests in `chatloop` used a 5ms real-time
budget and relied on wall-clock scheduling to fire the startup guard
timer before the first stream part arrived. On loaded CI runners the
timer sometimes lost the race, producing `attempts == 2` instead of
`attempts == 1` and flaking `TestRun_FirstPartDisarmsStartupTimeout`.
Replace the real `time.Timer` in `startupGuard` with a `quartz.Timer` so
tests can control time deterministically. Production behavior is
unchanged: `RunOptions.Clock` defaults to `quartz.NewReal()` when nil,
and the startup timeout still covers both opening the provider stream
and waiting for the first stream part.
- Add `RunOptions.Clock quartz.Clock` with nil-safe default.
- Tag the startup guard timer as `"startupGuard"` for quartz trap
targeting.
- Rewrite the four startup-timeout integration tests to use
`quartz.NewMock(t)` with trap/advance/release sequences instead of
wall-clock sleeps.
- Add `awaitRunResult` helper so tests fail with a clear message instead
of hanging when `Run` does not complete.
Closes https://github.com/coder/internal/issues/1460
Adds a GitHub Actions workflow that runs on PRs targeting `release/*`
branches to flag non-bug-fix cherry-picks.
## What it does
- Triggers on `pull_request_target` (opened, reopened, edited) for
`release/*` branches
- Checks if the PR title starts with `fix:` or `fix(scope):`
(conventional commit format)
- If not a bug fix, comments on the PR informing the author and emits a
warning (via `core.warning`), but does **not** fail the check
- Deduplicates comments on title edits by updating an existing comment
(identified by a hidden HTML marker) instead of creating a new one
> [!NOTE]
> Generated by Coder Agents
Adds `coder exp chat context add` and `coder exp chat context clear`
commands that run inside a workspace to manage chat context files via
the agent token.
`add` reads instruction and skill files from a directory (defaulting to
cwd) and inserts them as context-file messages into an active chat.
Multiple calls are additive — `instructionFromContextFiles` already
accumulates all context-file parts across messages.
`clear` soft-deletes all context-file messages, causing
`contextFileAgentID()` to return `!found` on the next turn, which
triggers `needsInstructionPersist=true` and re-fetches defaults from the
agent.
Both commands auto-detect the target chat via `CODER_CHAT_ID` (already
set by `agentproc` on chat-spawned processes), or fall back to
single-active-chat resolution for the agent. The `--chat` flag overrides
both.
Also adds sub-agent context inheritance: `createChildSubagentChat` now
copies parent context-file messages to child chats at spawn time, so
delegated sub-agents share the same instruction context without
independently re-fetching from the workspace agent.
<details><summary>Implementation details</summary>
**New files:**
- `cli/exp_chat.go` — CLI command tree under `coder exp chat context`
**Modified files:**
- `agent/agentcontextconfig/api.go` — `ConfigFromDir()` reads context
from an arbitrary directory without env vars
- `codersdk/agentsdk/agentsdk.go` — `AddChatContext`/`ClearChatContext`
SDK methods
- `coderd/workspaceagents.go` — POST/DELETE handlers on
`/workspaceagents/me/chat-context`
- `coderd/coderd.go` — Route registration
- `coderd/database/queries/chats.sql` — `GetActiveChatsByAgentID`,
`SoftDeleteContextFileMessages`
- `coderd/database/dbauthz/dbauthz.go` — RBAC implementations for new
queries
- `coderd/x/chatd/subagent.go` — `copyParentContextFiles` for sub-agent
inheritance
- `cli/root.go` — Register `chatCommand()` in `AGPLExperimental()`
**Auth pattern:** Uses `AgentAuth` (same as `coder external-auth`) —
agent token via `CODER_AGENT_TOKEN` + `CODER_AGENT_URL` env vars.
</details>
> 🤖 Generated by Coder Agents
---------
Co-authored-by: Michael Suchacz <203725896+ibetitsmike@users.noreply.github.com>
## Problem
`go run` caches the final linked executable in `~/.cache/go-build`.
Every
helper invocation via `go run ./scripts/<tool>` stores a copy, and
because
the cache key includes build metadata, the same tool accumulates
multiple
cached executables over time. With 12+ helper binaries invoked during
`make gen` and `make pre-commit`, this is a meaningful contributor to
GOCACHE growth.
## Fix
Replace `go run` with `go build -o _gen/bin/<tool>` for 12 repo-local
helper packages (16 Makefile callsites). Each helper is an explicit Make
file target with `$(wildcard *.go)` prerequisites, so `make -j`
serializes
builds correctly instead of racing on shared output paths.
Helpers converted: `apitypings`, `auditdocgen`, `check-scopes`,
`clidocgen`, `dbdump`, `examplegen`, `gensite`, `apikeyscopesgen`,
`metricsdocgen`, `metricsdocgen-scanner`, `modeloptionsgen`, `typegen`.
Left on `go run` (intentionally): `migrate-ci` and `migrate-test`
(CI/test-only, not on common developer paths).
`_gen/` is already in `.gitignore`. The `clean` target removes
`_gen/bin`.
## GOCACHE growth (isolated cache, single `make gen`)
| | Old (`go run`) | New (`go build -o`) |
|--|----------------|---------------------|
| Total cache size | 2.9 GB | 2.6 GB |
| Cached executables | 11 | 4 |
| Executable bytes | 401 MB | 25 MB |
The 4 remaining executables come from tools outside this change
(`dbgen` and `goimports` from `generate.sh`, plus two `main` binaries
from deferred helpers). Helper binaries now live in `_gen/bin/`
(581 MB, gitignored, cleaned by `make clean`).
## Build time benchmarks
**Source changed** (content hash invalidated, forces recompile):
| Helper | `go run` | `go build -o` + run | Overhead |
|--------|---------|---------------------|----------|
| typegen | 1.50s | 2.03s | +0.52s |
| examplegen | 1.37s | 1.67s | +0.30s |
| apikeyscopesgen | 1.21s | 1.71s | +0.50s |
| modeloptionsgen | 1.23s | 1.64s | +0.41s |
**Repeat invocation** (no source change, the common `make gen` / `make
pre-commit` path):
| Helper | `go run` (cache lookup) | Cached binary | Speedup |
|--------|------------------------|---------------|---------|
| typegen | 0.346s | 0.037s | 9.4x |
| examplegen | 0.368s | 0.037s | 9.9x |
| modeloptionsgen | 0.342s | 0.021s | 16.3x |
| apikeyscopesgen | 0.298s | 0.030s | 9.9x |
When source changes, `go build -o` is 0.3-0.5s slower per helper (it
writes a local binary instead of caching in GOCACHE). On repeat runs
(the common path), the pre-built binary is 10-16x faster because
`go run` still does a staleness check while the binary just executes.
> This PR was authored by Mux on behalf of Mike.
the current page has an "Agentic loop completed" block that doesn't
really contain any valuable info that isn't available elsewhere. replace
this with a status indicator
<img width="507" height="300" alt="Screenshot 2026-04-08 at 2 47 40 PM"
src="https://github.com/user-attachments/assets/09cf3772-a52d-485d-a15e-b2257b2d9003"
/>
couple of little design tweaks to make the UI of the Request Logs page
and Sessions pages more consistent:
- decrease size of Request Logs page chevron
- copy Request Logs page chevron animation for Sessions expandable
sections
- use TokenBadges component in RequestLogsRow
- wrap tool call counts in badges
<img width="1393" height="210" alt="Screenshot 2026-04-08 at 1 56 10 PM"
src="https://github.com/user-attachments/assets/97e7acb6-71c7-48d6-b0df-a102c7602cc0"
/>
Frontend for https://github.com/coder/coder/pull/24022.
From that PR's description:
> The agents chat interface displays thumbnails for videos recorded by
the computer use agent. Currently, to display a thumbnail, the frontend
downloads the entire video and shows the first frame.
#24022 adds a thumbnail file id to `wait_agent` tool results, and this
PR displays it instead of fetching the entire video.
The agents chat interface displays thumbnails for videos recorded by the
computer use agent. Currently, to display a thumbnail, the frontend
downloads the entire video and shows the first frame. This PR starts
storing a new thumbnail file in the database for every recorded video,
and exposes the file id in the `wait_agent` tool result alongside the
recording file id, so the frontend can fetch just the thumbnail.
The cherry-pick and backport workflows create PRs under
`github-actions[bot]`. Since GitHub doesn't support creating PRs on
behalf of another user, this adds attribution to the user who added the
label (`github.event.sender.login`):
- **Assignee**: the labeler is assigned to the backport PR
- **Reviewer**: the labeler is added as a reviewer
- **PR body**: includes "Requested by: @user"
Applied to both `cherry-pick.yaml` and `backport.yaml`.
---
> Generated by Coder Agents
## Motivation
During the April 2 dogfood incident, a pod OOM-kill triggered a
reconnection storm: hundreds of chat-stream and agent-RPC websockets all
attempted to reconnect at the same deterministic backoff intervals (1 s,
2 s, 4 s, …). Because every browser tab computed the same delay, the
surviving replicas received a synchronized wall of new connections at
each retry tick, amplifying the overload that caused the first OOM in
the first place.
The root cause of the memory blowup (chatd serialization cost) is a
separate issue. This change addresses the secondary blast-radius
problem: when N clients reconnect in lockstep, the retry storm itself
becomes a capacity threat.
## Change
The shared `createReconnectingWebSocket` utility now applies symmetric
jitter (default ±30%) to the capped exponential-backoff delay before
scheduling the reconnect timer. With 100 clients and a 1 s base delay,
reconnects spread over the 700 ms–1300 ms window instead of all landing
at exactly 1000 ms, and once retries hit `maxMs` the scheduler still
preserves downward spread instead of collapsing back to a single tick.
Two new options are accepted by callers:
- **`jitter`** (0–1 fraction, default `0.3`) — controls the jitter
window. Values are clamped to `[0, 1]`; `0` preserves exact legacy
timing.
- **`random`** (`() => number`, default `Math.random`) — injectable RNG,
primarily a deterministic test seam. Non-finite output falls back to the
midpoint (`0.5`).
The `retryingAt` timestamp surfaced to `ChatStatusCallout` is computed
from the jittered delay, so the countdown shown to users reflects the
actual retry time. The scheduler also keeps `maxMs` as a hard ceiling on
the final delay and saturates exponential overflow at that cap instead
of dropping to `0ms` retries.
No production callers need changes — the default jitter activates
automatically for all four call sites (`AgentsPage` chat-list watcher,
`AgentChatPage` workspace watcher, `useChatStore` per-chat stream,
`useGitWatcher`). The two downstream tests that asserted exact reconnect
timing now pin `Math.random()` to `0.5` so those expectations stay
deterministic.
This pull-request makes a few changes to our `<Badge />` component to
bring it inline with Figma.
* Added all variants to the stories of Figma (they can vary per
badge-type, so its better we track everything).
* Removed the `border` variant of the component, border variants should
be on all `sm` and `md`.
* Added a hover effect to the `default` variant (per-design).
* Resolved issue with sizings of `xs` and `sm` plus resolved
iconography.
* Resolved issue with icons not showing at all on `xs` variants.
closes CODAGT-122
Add a spacer div that renders only when an assistant message lacks the
action bar, matching the height the action bar would provide.
> 🤖 Generated by Coder Agents
## Problem
Workspaces showed as "Healthy" immediately after creation while the
agent was still downloading, starting, or connecting. If the agent never
connected, the workspace stayed "Healthy" for the entire connection
timeout (~120s), then abruptly flipped to "Unhealthy".
## Root cause
In `db2sdk.WorkspaceAgent`, the health switch had no case for
`WorkspaceAgentConnecting`. Agents in `connecting` status with a
non-`off` lifecycle (e.g. `created` after a fresh build) fell through to
the `default` case and were marked `Healthy = true`.
## Fix
Add an explicit case for `WorkspaceAgentConnecting` that sets `Healthy =
false` with reason `"agent has not yet connected"`. The case is placed
after the existing `!connected + off` case (which correctly catches
stopped agents as "not running") and before the `timeout`/`disconnected`
cases.
```
Status + Lifecycle → Health reason
──────────────────────────────────────────────────────
any !connected + off → "agent is not running"
connecting + created/starting → "agent has not yet connected" ← NEW
timeout + any → "agent is taking too long to connect"
disconnected + any → "agent has lost connection"
connected + start_error → "agent startup script exited with an error"
connected + shutting_down → "agent is shutting down"
connected + ready/starting → healthy
```
The frontend already handles this case — `getAgentHealthIssue()` returns
"Workspace agent is still connecting" with `severity: "info"` for
unhealthy workspaces with connecting agents.
## Test changes
- **Healthy test**: now actually connects the agent via `agenttest.New`
before asserting health (previously passed due to the bug).
- **New Connecting test**: verifies a never-connected agent is correctly
marked unhealthy.
- **Mixed health test**: connects a1 and waits for the mixed state
(`a1.Healthy && !workspace.Healthy`) to avoid a race where both agents
are initially connecting.
- **Sub-agent excluded test**: connects the parent agent and waits for
it to be healthy before creating the sub-agent.
- **TestWorkspaceAgent/Connect**: flipped assertion to `Health.Healthy
== false` for a `dbfake` agent that never connects.
<details>
<summary>Review notes</summary>
### Known follow-up
The `healthy:false` workspace search filter maps to `[disconnected,
timeout]` and does not include `connecting`. This is a pre-existing gap
that is now more consequential — a workspace unhealthy solely due to a
connecting agent won't appear in `healthy:false` results. Worth a
follow-up issue.
### Deep review findings addressed
| Finding | Severity | Status |
|---------|----------|--------|
| Mixed health test race (all 3 reviewers) | P2 | Fixed — tightened
`Eventually` condition |
| `TestWorkspaceAgent/Connect` assertion break | P1 | Fixed — flipped
assertion |
| CLI renders red for connecting agents | Obs | Acknowledged — design
trade-off, accurate but visually strong for transient state |
| Switch case ordering overlap | Obs | Documented with inline comment |
</details>
> 🤖 This PR was created with the help of Coder Agents, and needs a human
review. 🧑💻
Replace Tooltip with `HelpPopover` in the "New workspace" page header.
`HelpPopover` supports interactive content like links and provides
better layout control, making it a better fit for this use case.
Renames the "Security implications" section to "Security posture" and
reframes the intro paragraph. "Implications" reads as a caveat or
warning; the section actually describes built-in structural guarantees
of the control plane architecture.
> PR generated with Coder Agents
Workspace agent logs could still fail after the earlier invalid UTF-8
fix because NUL bytes are valid Go/protobuf strings but are rejected by
Postgres text columns. The legacy HTTP log upload path also bypassed the
old sanitization entirely, and both server insert paths computed
logs_length from the unsanitized input.
Add a shared log-output sanitizer in agentsdk, use it in the protobuf
conversion path and both server-side insert paths, and compute
OutputLength from the sanitized string so overflow accounting matches
what is actually stored. This keeps the old invalid UTF-8 behavior while
also handling embedded NUL bytes consistently across DRPC and HTTP log
ingestion.
Refs [#23292 ](https://github.com/coder/coder/issues/23292)
Refs [#13433 ](https://github.com/coder/coder/issues/13433)
Adds backend validation for user secret environment variable names and file paths.
Env name validation enforces POSIX naming rules and blocks a deliberately aggressive denylist of reserved names and prefixes. The denylist errs on the side of blocking too much since it's easier to remove entries later than to add them after users have created conflicting secrets.
File path validation requires paths to start with ~/ or /.
Fixes several documentation gaps and inaccuracies in the Coder Agents
docs identified during a deep review against the current product state.
## BYOK (User API Keys)
`models.md` stated *"Developers cannot add their own providers, models,
or API keys"* — this has been incorrect since the provider key policy
system shipped (Apr 2, #23751/#23781).
- Added **Key policy** section documenting the three admin toggles
(`central_api_key_enabled`, `allow_user_api_key`,
`allow_central_api_key_fallback`) with a truth table showing all
resolution outcomes
- Added **User API keys (BYOK)** section covering the developer-facing
key management page, status indicators, selection priority, and key
removal
- Updated `platform-controls/index.md` to reference BYOK instead of
claiming keys are admin-only
## Reasoning effort enum fixes
- **OpenAI**: removed `none` — code accepts `minimal, low, medium, high,
xhigh`
- **OpenRouter**: narrowed to `low, medium, high` per
`ReasoningEffortFromChat` in `chatprovider.go`
## Tool table completeness
- Added `spawn_computer_use_agent`, `read_skill`, `read_skill_file` to
`index.md` tool table
- Added "Workspace extension tools" section to `architecture.md` for
`read_skill`/`read_skill_file`
- Fixed orchestration restriction note to list all 5 gated tools instead
of just `spawn_agent`
- Added conditional availability notes for desktop and skills tools
## Platform controls
Three admin-only settings existed in the Behavior tab with no
documentation:
- **Virtual desktop** — admin toggle, Anthropic + portabledesktop
requirements
- **Workspace autostop fallback** — default TTL for agent workspaces
without template-defined autostop
- **Data retention** — moved `chat-retention.md` into
`platform-controls/` since it's admin-only, fixed nav path
---
> PR generated with Coder Agents
Dev and RC builds now show diagonal warning stripes in the navbar plus a
centered version badge, making it impossible to miss which build you're
running.
**Devel build:** amber "warning" from theme
**RC build:** sky "pending" from theme
> 🤖 Written by a Coder Agent. Will be reviewed by a human.
## Summary
Adds `credential_kind` and `credential_hint` columns to
`aibridge_interceptions` to record how each LLM request was
authenticated and provide a masked credential identifier for audit
purposes.
This enables admins to distinguish between centralized API keys,
personal API keys, and subscription-based credentials in the
interceptions audit log.
## Changes
- New migration adding `credential_kind`and `credential_hint` to
`aibridge_interceptions`
- Updated `InsertAIBridgeInterception` query and proto definition to
carry the new fields
- Wired proto fields through `translator.go` and `aibridgedserver.go` to
the database
Depends on https://github.com/coder/aibridge/pull/239
Update the release calendar table now that v2.31.7 has been promoted to
stable (`latest` on GitHub Releases).
## Changes
| Release | Old Status | New Status | Latest Patch |
|---------|-----------|------------|-------------|
| 2.31 | Mainline | Stable | v2.31.7 |
| 2.30 | Stable | Security Support | v2.30.6 |
| 2.29 | Security Support + ESR | Extended Support Release | v2.29.9 |
---
> **Note:** The auto-generation script
(`scripts/update-release-calendar.sh`) determines status positionally
from the latest non-RC tag, so it will always mark the latest minor
version as "Mainline". This manual update is needed to reflect the
promotion of 2.31 to stable.
Co-authored-by: blink-so[bot] <211532188+blink-so[bot]@users.noreply.github.com>
Adds an optional `CreatedAt` timestamp to `tool-call` and `tool-result`
`ChatMessagePart` variants so the frontend can compute tool execution
duration (`result.created_at - call.created_at`).
Timestamps are recorded at the correct moments in the chatloop:
- **Tool-call**: when the model stream emits the tool call
- **Tool-result**: when tool execution completes (or is interrupted)
These are passed through `PersistedStep.PartCreatedAt` so the
persistence layer can apply accurate timestamps to stored parts.
SSE-published parts also carry `CreatedAt` for real-time display.
Old persisted messages without `created_at` deserialize to `nil` — fully
backward compatible.
<details><summary>Implementation notes (Coder Agents
generated)</summary>
### Why not stamp in `PartFromContent`?
`PartFromContent` is called both for SSE publishing (correct timing) and
during persistence (wrong timing — both tool-call and tool-result would
get the same "persistence time" timestamp, yielding ~0 duration).
Instead, timestamps are captured in the chatloop at the right moments
and carried through `PersistedStep.PartCreatedAt` as a
`map[string]time.Time` keyed by `"call:<id>"` / `"result:<id>"`.
### Interrupted tool calls
`persistInterruptedStep` also stamps `CreatedAt` on synthetic error
results for cancelled/interrupted tool calls, so partial duration is
available.
### Files changed
| File | Change |
|------|--------|
| `codersdk/chats.go` | Add `CreatedAt *time.Time` field |
| `codersdk/chats_test.go` | JSON round-trip test |
| `coderd/database/dbtime/dbtime.go` | Add `TimePtr` helper |
| `coderd/x/chatd/chatloop/chatloop.go` | Track timestamps, pass through
`PersistedStep` |
| `coderd/x/chatd/chatd.go` | Apply timestamps during persistence |
| `coderd/x/chatd/chatprompt/chatprompt_test.go` | Verify
`PartFromContent` does NOT stamp |
| `site/src/api/typesGenerated.ts` | Auto-generated |
</details>
---------
Co-authored-by: Ethan <39577870+ethanndickson@users.noreply.github.com>
## What
Two small docs improvements for AI Bridge:
1. **`setup.md` – Structured Logging section**: Added a `record_type`
table documenting the six event types emitted by AI Bridge structured
logs (`interception_start`, `interception_end`, `token_usage`,
`prompt_usage`, `tool_usage`, `model_thought`) along with their key
fields. Previously only the `"interception log"` message prefix was
mentioned.
2. **`monitoring.md`**: Added a "Structured Logging" section that
cross-links to `setup.md#structured-logging`, so users landing on the
monitoring page can discover the feature without navigating to the setup
guide first.
<details><summary>Source reference</summary>
Record types and fields were extracted from
`enterprise/aibridgedserver/aibridgedserver.go` where they are emitted
as `slog.F("record_type", "...")` string literals under the
`InterceptionLogMarker` (`"interception log"`) message.
</details>
Adds client-executed dynamic tools to the chat API. Dynamic tools are
declared by the client at chat creation time, presented to the LLM
alongside built-in tools, but executed by the client rather than chatd.
This enables external systems (Slack bots, IDE extensions, Discord bots,
CI/CD integrations) to plug custom tools into the LLM chat loop without
modifying chatd's built-in tool set.
Modeled after OpenAI's Assistants API: the chat pauses with
`requires_action` status when the LLM calls a dynamic tool, the client
POSTs results back via `POST /chats/{id}/tool-results`, and the chat
resumes.
See [this example](https://github.com/coder/coder-slackbot-poc) as a
reference for how this is used. It's highly-configurable, which would
enable creating chats from webhooks, periodically polling, or running as
a Slackbot.
<details>
<summary>Design context</summary>
### Architecture
The chatloop **exits** when it encounters dynamic tools and
**re-enters** when results arrive. No blocking channels, no pubsub for
tool results, no in-memory registry. The DB is the only coordination
mechanism.
```
Phase 1 (chatloop):
LLM response → execute built-in tools only →
Persist(assistant + built-in results) →
status = requires_action → chatloop exits
Phase 2 (POST /tool-results):
Persist(dynamic tool results) →
status = pending → wakeCh → chatloop re-enters
```
### Validation (POST /tool-results)
1. Chat status must be `requires_action` (409 if not)
2. Read chat's `dynamic_tools` → set of dynamic tool names
3. Read last assistant message → extract tool-call parts matching
dynamic tool names
4. Submitted tool_call_ids must match exactly (400 for missing/extra)
5. Persist tool-result message parts, set status to `pending`, signal
wake
### Idempotency
Tool call IDs scoped per LLM step. State machine (`requires_action` →
`pending`) is the guard. First POST wins, subsequent get 409.
### Mixed tool calls
When the LLM calls both built-in and dynamic tools in one step, built-in
tools execute immediately. Their results are persisted in phase 1.
Dynamic tool results arrive via POST in phase 2. The LLM sees all
results when the chatloop resumes.
</details>
> 🤖 Generated by Coder Agents
## Problem
In linked worktrees, Git hooks inherit multiple repo-local environment
variables: `GIT_DIR`, `GIT_COMMON_DIR`, `GIT_INDEX_FILE`, and others.
The
pre-commit and pre-push hooks only unset `GIT_DIR`, leaving the rest in
place.
When `make pre-commit` runs `go build`, Go tries to stamp VCS info by
shelling
out to `git`. With the leftover partial Git environment, `git` exits 128
and
the build fails:
```
error obtaining VCS status: exit status 128
Use -buildvcs=false to disable VCS stamping.
```
This only happens inside hooks in a linked worktree — running `make
pre-commit`
directly from the terminal works fine because the repo-local vars are
not set.
## Fix
Replace the bare `unset GIT_DIR` in both hooks with a loop that clears
every
variable reported by `git rev-parse --local-env-vars`:
```sh
while IFS= read -r var; do
unset "$var"
done < <(git rev-parse --local-env-vars)
```
This covers all 15 repo-local variables Git may inject (`GIT_DIR`,
`GIT_COMMON_DIR`, `GIT_INDEX_FILE`, `GIT_OBJECT_DIRECTORY`, etc.) and is
forward-compatible — if Git adds new local vars in the future, the loop
picks
them up automatically.
Faur Ioan-Aurel