fix(dm+engine): the 2 opus golden-spine RED bugs (OOC seating-leak + CompanionQuestArc.summary) - #1077
Merged
Merged
Conversation
…CompanionQuestArc.summary)
Both surfaced on the gs-ember-opus golden-spine (embergloom-pact, 24 beats, OPUS DM)
that RED'd on narration_no_ooc_leak + no_rejected_tool_calls.
1a — narration_no_ooc_leak: the DM narrated its own seating decision into player prose
("Now I'll seat Bryn as the player character — a hedge-healer's daughter fits a Bard…")
on the opening beat. The FICTION-ONLY rule (SKILL.md:83) already bans the phrasing, but the
leak keeps shipping on the very first beat where seating happens. Add a sharp wrong→correct
'SEAT THE PC SILENTLY' reminder anchored to the cold-open pointer (no engine change).
1b — no_rejected_tool_calls: set_companion_quest_arc(arc={"summary": ...}) was REJECTED
because CompanionQuestArc (extra='forbid') had no summary field (extra_forbidden ⇒ FATAL ⇒
lenses RED-capped). Add summary: str = '' additively (mirrors the #1048 per-stage description
fix) + 2 regression tests (accepted+persisted; old snapshots round-trip). 77/77 pass.
Refs the standardized-closeout opus run; un-RED-caps authored-companion-arc golden-spines.
📝 WalkthroughWalkthroughAdds a ChangesCompanionQuestArc summary field
DM Beat Cycle Silent Seating Guidance
Estimated code review effort🎯 1 (Trivial) | ⏱️ ~5 minutes Possibly related PRs
Poem
🚥 Pre-merge checks | ✅ 4 | ❌ 1❌ Failed checks (1 warning)
✅ Passed checks (4 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches📝 Generate docstrings
Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 1
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@servers/engine/tests/test_companion_arc.py`:
- Around line 516-522: The test function
test_companion_quest_arc_summary_defaults_empty_and_old_snapshots_round_trip
claims to verify round-trip behavior but only checks the immediate response from
set_companion_quest_arc without verifying persistence. To fix this, after the
existing assertion on the set_companion_quest_arc response, add a call to
get_companion_quest_arcs to retrieve the persisted companion quest arc data and
assert that the summary field in the retrieved data is still an empty string,
ensuring the data actually round-trips correctly through storage and retrieval.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Organization UI
Review profile: CHILL
Plan: Pro Plus
Run ID: a5c60680-cb45-4b6a-8367-536f0a2248eb
📒 Files selected for processing (3)
servers/engine/models.pyservers/engine/tests/test_companion_arc.pyskills/dungeon-master/SKILL.md
Comment on lines
+516
to
+522
| def test_companion_quest_arc_summary_defaults_empty_and_old_snapshots_round_trip(camp): | ||
| # Additive invariant: an arc authored WITHOUT a summary keeps today's behavior (empty string), | ||
| # so pre-fix snapshots deserialize unchanged. | ||
| cid, comp = camp | ||
| out = server.set_companion_quest_arc(cid, comp, {"title": "No Summary Arc", "status": "locked"}) | ||
| assert out["companion_quest_arc"]["summary"] == "" | ||
|
|
There was a problem hiding this comment.
Test name claims round-trip but only verifies write response.
This test currently checks the immediate set_companion_quest_arc payload, not persisted read-back. Add a get_companion_quest_arcs assertion so the “round trip” claim is actually covered.
Suggested patch
def test_companion_quest_arc_summary_defaults_empty_and_old_snapshots_round_trip(camp):
@@
cid, comp = camp
out = server.set_companion_quest_arc(cid, comp, {"title": "No Summary Arc", "status": "locked"})
assert out["companion_quest_arc"]["summary"] == ""
+ got = server.get_companion_quest_arcs(cid, companion_id=comp)["companion_quest_arcs"][0]
+ assert got["summary"] == ""📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
Suggested change
| def test_companion_quest_arc_summary_defaults_empty_and_old_snapshots_round_trip(camp): | |
| # Additive invariant: an arc authored WITHOUT a summary keeps today's behavior (empty string), | |
| # so pre-fix snapshots deserialize unchanged. | |
| cid, comp = camp | |
| out = server.set_companion_quest_arc(cid, comp, {"title": "No Summary Arc", "status": "locked"}) | |
| assert out["companion_quest_arc"]["summary"] == "" | |
| def test_companion_quest_arc_summary_defaults_empty_and_old_snapshots_round_trip(camp): | |
| # Additive invariant: an arc authored WITHOUT a summary keeps today's behavior (empty string), | |
| # so pre-fix snapshots deserialize unchanged. | |
| cid, comp = camp | |
| out = server.set_companion_quest_arc(cid, comp, {"title": "No Summary Arc", "status": "locked"}) | |
| assert out["companion_quest_arc"]["summary"] == "" | |
| got = server.get_companion_quest_arcs(cid, companion_id=comp)["companion_quest_arcs"][0] | |
| assert got["summary"] == "" |
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In `@servers/engine/tests/test_companion_arc.py` around lines 516 - 522, The test
function
test_companion_quest_arc_summary_defaults_empty_and_old_snapshots_round_trip
claims to verify round-trip behavior but only checks the immediate response from
set_companion_quest_arc without verifying persistence. To fix this, after the
existing assertion on the set_companion_quest_arc response, add a call to
get_companion_quest_arcs to retrieve the persisted companion quest arc data and
assert that the summary field in the retrieved data is still an empty string,
ensuring the data actually round-trips correctly through storage and retrieval.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What
Fixes the two behavioral-gate REDs the OPUS golden-spine (
gs-ember-opus, embergloom-pact, 24 beats) surfaced — the decisive run that proved opus drives the campaign (act 2, 3 locations, quest resolved) and scored story 4.9 (the "DM under-drives" scare was a Sonnet-measurement artifact, not a regression).1a ·
narration_no_ooc_leak— OOC seating-preamble leak (DM prompt)The DM narrated its own casting decision into player-facing prose on the opening beat:
The FICTION-ONLY rule (
SKILL.md:83) already bans this exact family, but it keeps shipping on the very first beat where PC-seating happens. Fix: a sharp wrong→correct "SEAT THE PC SILENTLY" reminder anchored right at the cold-open pointer. No engine change.1b ·
no_rejected_tool_calls—CompanionQuestArc.summaryrejected (engine model)set_companion_quest_arc(arc={"summary": ...})was REJECTED —CompanionQuestArcisextra="forbid"and had nosummaryfield (extra_forbidden→no_rejected_tool_callsFATAL → all lenses RED-capped). Addsummary: str = ""additively (mirrors the just-merged #1048 per-stagedescriptionfix).Why it matters
Both are the cap-rate failure class — a single rejected tool call / OOC leak RED-caps the whole run's lenses, which is what kept the honest opus story score (4.9 raw) from showing as a clean GREEN run.
Tests
+2regression tests intest_companion_arc.py(summary accepted+persisted; arc without summary defaults""so old snapshots round-trip).uv run --directory servers/engine python -m pytest tests/test_companion_arc.py→ 77 passed.Invariants
Additive-by-default (empty == today; old snapshots round-trip); engine stays sole-writer; no gate weakened. 1a is a prompt-only change.
Verification (post-merge)
A short opus golden-spine →
narration_no_ooc_leakGREEN (seating leak gone). Reported in the standardized closeout.Summary by CodeRabbit
New Features
Documentation
Tests