feat(tasks): replace turn_complete thread writeback with pr_created artifact messages

[k11kirky]

Problem

Task threads (the PostHog Code thread panel) mixed agent turn writebacks with human conversation. Threads should be a human-to-human surface, with the agent represented by the artifacts it produced. Canvases already announce themselves (canvas_created), but pull requests never did — and every background turn ended with a turn_complete agent message in the thread.

Changes

• Removed post_turn_complete_thread_update and its call from the sandbox event relay (plus the now-dead final-message tracking in the relay loop). Background runs no longer write the agent's closing message into the thread.
• list_thread_messages excludes legacy turn_complete rows so previously written turn messages don't resurface in clients.
• New post_pr_created_thread_update: when a run's PR URL is recorded, an agent artifact message (event="pr_created", payload.pr_url, content [owner/repo#N](url) has been opened) lands in the task thread. Called from the run output paths (set_task_run_output, update_task_run) and the GitHub webhook backstop (_record_run_pr_url), deduped per PR URL via the task's existing pr_created rows. Same feature-flag and channel guards as the other agent thread updates.

The client side (rendering pr_created/canvas_created as artifact rows and dropping agent turn/prompt rows) is PostHog/code#3694.

How did you test this code?

Red/green: rewrote products/tasks/backend/tests/test_thread_updates.py first (failing on the missing function), then implemented. Tests I (Claude) actually ran locally, all green:

• test_thread_updates (13 tests): pr_created content/payload, per-URL dedupe, distinct-PR handling, flag/channel/creator skip guards, canvas_created regression, turn_complete exclusion from listing
• test_channels_api + test_webhooks (100 tests)
• test_facade
• test_relay_sandbox_events (53 passed; one pre-existing test errored in my sandbox because the persons-DB sqlx fixture isn't available there — unrelated to this diff)
• ruff check / ruff format clean

Automatic notifications

• Publish to changelog?
• Alert Sales and Marketing teams?

Docs update

🤖 Agent context

Autonomy: Fully autonomous

Authored by Claude (PostHog Code cloud task) together with the paired client PR. Followed red-green testing: wrote the failing tests before each implementation change. Chose DB-backed dedupe (existing pr_created rows per URL) over a Redis cooldown since the agent-output path and webhook backstop can fire arbitrarily far apart.

---

Created with PostHog Code

[github-actions]

Hey @k11kirky! 👋

It looks like your git author email on this PR isn't your @posthog.com address (k11kirky@gmail.com). Since you're on the PostHog team, it's worth pointing your local git author email at your @posthog.com address. Why it matters:

• Consistent work identity in git history — internal tooling that attributes commits to team members keys off your @posthog.com address.
• Keeps team contributions easy to tell apart from external community ones when scanning history.

You can fix it for this repo with:

git config user.email "you@posthog.com"

Or set it globally with git config --global user.email "you@posthog.com". No need to redo this PR — just a nudge for next time. 🙂

[greptile-apps]

Prompt To Fix All With AI

Fix the following 2 code review issues. Work through them one at a time, proposing concise fixes.

---

### Issue 1 of 2
products/tasks/backend/facade/api.py:5501-5506
**PR Deduplication Races**

The agent-output path and GitHub webhook can both pass this `exists()` check before either creates its message. Because there is no matching database uniqueness constraint or lock around the check-and-create sequence, the same PR can appear twice in the task thread.

### Issue 2 of 2
products/tasks/backend/facade/api.py:5261-5263
**Legacy Mentions Target Hidden Messages**

Legacy `turn_complete` messages created mention records for task creators, but this filter only hides the messages from the thread API. Those records remain in the separate mentions feed, so opening an existing notification leads to a thread where the referenced message is missing.

_{Reviews (1): Last reviewed commit: "feat(tasks): replace turn_complete threa..." | Re-trigger Greptile}

[veria-ai]

PR overview

All previously flagged issues have been addressed. No open security concerns remain on this pull request.

Security review

No open security issues remain on this pull request.

Fixed/addressed: 2 · PR risk: 0/10

[tatoalo]

Could we ensure the announcement even when pr_url is already populated? _record_run_output_field returns False and webhook retries can never repair the missing artifact. It would be safer to idempotently ensure the artifact whenever the persisted URL matches this webhooks one.

Agree also on greptile + veria points, thanks for this!

[github-actions]

🤖 CI report

⚠️ Backend coverage — 98.0% of changed backend lines covered — 2 uncovered

🧪 Backend test coverage

Patch coverage — changed backend lines (products + core): ████████████████████ 98.0% (120 / 122)

File	Patch	Uncovered changed lines
products/tasks/backend/facade/api.py	94.4%	5360, 5405

🤖 Agents: add a test covering the lines above, or note why under "How did you test this code?". Machine-readable gap list: the patch-coverage artifact on this run (gh run download 30046846386 -n patch-coverage), or the coverage-data block at the end of this comment.

Per-product line coverage (touched products)

Product	Coverage	Lines
demo	███████████░░░░░░░░░ 56.2%	1,497 / 2,663
tasks	██████████████░░░░░░ 69.4%	30,037 / 43,278
signals	████████████████░░░░ 79.9%	21,906 / 27,415
cdp	████████████████░░░░ 81.0%	3,144 / 3,881
data_modeling	█████████████████░░░ 82.7%	5,547 / 6,708
notebooks	█████████████████░░░ 85.3%	7,275 / 8,531
agent_platform	█████████████████░░░ 86.4%	3,807 / 4,405
actions	█████████████████░░░ 86.6%	717 / 828
cohorts	██████████████████░░ 87.8%	4,488 / 5,114
product_tours	██████████████████░░ 87.9%	1,303 / 1,482
exports	██████████████████░░ 88.4%	6,949 / 7,861
data_warehouse	██████████████████░░ 88.9%	11,887 / 13,378
conversations	██████████████████░░ 89.3%	16,834 / 18,848
engineering_analytics	██████████████████░░ 89.4%	6,386 / 7,145
dashboards	██████████████████░░ 89.4%	5,983 / 6,693
error_tracking	██████████████████░░ 89.7%	10,174 / 11,342
alerts	██████████████████░░ 89.9%	4,054 / 4,508
early_access_features	██████████████████░░ 90.1%	1,031 / 1,144
mcp_analytics	██████████████████░░ 90.1%	2,763 / 3,065
streamlit_apps	██████████████████░░ 90.4%	2,501 / 2,767
slack_app	██████████████████░░ 90.7%	9,028 / 9,951
marketing_analytics	██████████████████░░ 91.0%	11,792 / 12,964
stamphog	██████████████████░░ 91.1%	4,056 / 4,450
product_analytics	███████████████████░ 92.7%	5,811 / 6,268
ai_observability	███████████████████░ 92.7%	14,995 / 16,168
surveys	███████████████████░ 93.0%	5,730 / 6,163
web_analytics	███████████████████░ 93.1%	14,398 / 15,461
posthog_ai	███████████████████░ 93.2%	1,326 / 1,422
approvals	███████████████████░ 93.3%	3,437 / 3,682
reminders	███████████████████░ 93.4%	468 / 501
workflows	███████████████████░ 93.6%	6,390 / 6,826
endpoints	███████████████████░ 94.1%	8,640 / 9,177
revenue_analytics	███████████████████░ 94.6%	3,651 / 3,858
review_hog	███████████████████░ 94.7%	6,814 / 7,199
skills	███████████████████░ 94.7%	3,086 / 3,258
logs	███████████████████░ 95.4%	9,935 / 10,416
experiments	███████████████████░ 95.7%	25,036 / 26,159
replay_vision	███████████████████░ 96.2%	15,170 / 15,774
annotations	███████████████████░ 96.2%	732 / 761
feature_flags	███████████████████░ 96.5%	17,110 / 17,734
user_interviews	███████████████████░ 96.5%	2,638 / 2,734
customer_analytics	███████████████████░ 97.1%	9,047 / 9,317
warehouse_sources	███████████████████░ 97.1%	330,211 / 339,939
data_catalog	███████████████████░ 97.4%	2,367 / 2,429
pulse	████████████████████ 98.4%	2,017 / 2,049

Report-only. Patch coverage = changed backend lines covered vs origin/master. Sorted lowest first.
Known gaps: lines covered only by Temporal tests show as uncovered; core line numbers may drift if master changed the same file.

[posthog-bot-comment-resolver]

🔀 Tried to auto-resolve conflicts with master but this one needs a human.

I won't retry until the branch or master moves.

[deployment-status-posthog]

Deploy status

Environment	Status	Deployed At	Workflow
dev	✅ Deployed	2026-07-23 22:52 UTC	Run
prod-us	✅ Deployed	2026-07-23 23:08 UTC	Run
prod-eu	✅ Deployed	2026-07-23 23:08 UTC	Run