fix(naming): finish the dowel rename in this repo, and mark what stays #1005
No reviewers
Labels
No labels
move-to-repo
coilyco-bridge-deploy
move-to-repo
coilyco-flight-deck-agent-compose
move-to-repo
coilyco-gaming-eco-app
move-to-repo
coilysiren-inbox
move-to-repo
unknown
🔒⚠️📦⚠️🔒 SANDBOXED 🔒⚠️📦⚠️🔒
autonomy
async-consult
autonomy
epic
autonomy
headless
autonomy
live-collab
c#
priority
P0
priority
P1
priority
P2
priority
P3
priority
P4
role/ai
role/creator
role/design
role/director
role/engineer
role/exec
role/human
role/ops
role/qa
No milestone
No project
No assignees
1 participant
Notifications
Due date
No due date set.
Dependencies
No dependencies set
Reference
coilyco-gaming/sirens-echo!1005
Loading…
Reference in a new issue
No description provided.
Delete branch "aos/claude/finish-dowel-rename"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Closes #948. Answers ask 4 of coilyco-bridge/deploy#657.
Two corrections to #948 before the diff
There are six references across four files, not three. #948 named
docs/sirens-echo-http.mdandinternal/community/silentturn_test.go. It missedAGENTS.mdand the two in.agents/skills/sirens-dowel/references/moxn-knowledge-base.md.None of them is a test name. #948's case for urgency rested on one being a test name, invoking the
TestTheCadenceIsEightSixteenFortyprecedent where a name stops describing the thing and passes anyway. The two test functions in that file areTestATurnThatAlreadyAnsweredThroughAToolPostsNothingandTestATurnThatDidNothingAtAllIsStillAFailure. Neither carries the lane, and neither does the fixture type. The reference is in a comment above the fixture. Real, worth fixing, not the failure mode described.What changed, and why each one
#948 said to read deploy#657 first, because some references there are deliberate survivors. I did, and the split in this repo is cleaner than in deploy.
docs/sirens-echo-http.md, two references — factually wrong, fixed. The cluster has cut over: the namespace issirens-dowel, andsirens-deep-owl-glassno longer exists in the manifests or on kai-server. The doc named that namespace twice. I checked the surrounding facts againstsirens-dowel-values.yamlrather than assuming they travelled with the name, and both still hold: NodePort 30122 andtailnet.enabled: false.AGENTS.md— retired name, fixed. It called the lane "the owl.glass lane" while naming its skill rootsirens-dowel.internal/community/silentturn_test.go— retired name, rewritten rather than substituted. The comment called the fixture "the owl.glass shape". What it describes is not specific to that lane at all: it is the shape any write tool aimed at the reply channel produces. It now says that.I first wrote a version citing an issue number for the record behind it, then checked and the commit that added the file (
c0973a5) names no issue, so the citation came out rather than going in unverified..agents/skills/sirens-dowel/references/moxn-knowledge-base.md, two references — deliberate survivors, kept and annotated.owl-glassthere is the Moxn workspace, andowl-glass.moxn.devis its address. Moxn's names, not this deployment's, and they did not rename with the lane. Renaming them would have broken a working reference.deploy#657's actual complaint is that "nothing distinguishes a deliberate survivor from a missed one," so these now carry one line saying why they stay. That is the discipline its ask 1 requests, applied to this repo's two.
Not in scope, still open on deploy#657
The telemetry discontinuity is the part that bites tomorrow and is not mine to fix here.
service.namemoved with the namespace, so every SigNoz query keyed tosirens-deep-owl-glassnow returns empty rather than erroring, and the #939/#940 incident measurements are all against the old name. Anyone asking "did the fix work" on August 19 has to union two service names or read an empty result as a pass. deploy#657 ask 3 carries it.Also untouched, correctly: the incident issues themselves still say
owl.glass, per Darren's reasoning on #657 that a record silently adopting a name the incident did not happen under is harder to trust.Verification
just gatepasses: build, policy-check, vet, test, test-skips, pre-commit. Naming only, no behaviour, and the one Go file touched is a comment.