coilyco-flight-deck/ward

Fork 0

Audit old ward-salvage branches for recoverable work #666

New issue

Closed

opened 2026-07-08 08:39:37 +00:00 by coilyco-ops · 2 comments

coilyco-ops commented

2026-07-08 08:39:37 +00:00

Member

Goal

Audit the old ward-salvage/* branches that still contain real branch-only changes and determine whether each holds recoverable lost work, has already landed through another path, or is obsolete/superseded.

Branches to inspect

ward-salvage/ward-0d7db064 - branch-only files suggest error reporting work.
ward-salvage/ward-33ebc07c - workflow mirror files/docs.
ward-salvage/ward-71d4c8f9 - agent reap implementation files.
ward-salvage/ward-86ffb40e - agent adapter/fleet test changes.
ward-salvage/ward-9b76e536 - director consult files/docs.
ward-salvage/ward-df8343fa - graft-removal doc differs from main.
ward-salvage/ward-e6ca4924 - advisor/reply/doc changes.

Context

A director surface audit found five other salvage branches are provenance-only and tied to already closed issues: ward-11906072, ward-9686d657, ward-9c289e1b, ward-d9e51395, and ward-7ca66e36. Do not spend time on those unless needed as comparison.

The seven branches above are concerning because their branch-only diffs include real code/doc changes from about a week ago. Some may correspond to closed issues whose work later landed differently. Others may be lost work that should be recovered.

Do

For each branch:

Fetch the branch and compare it to current origin/main.
Identify the likely source issue from commit messages, file names, and issue search.
Decide one of:
- recover - useful work is not on main and should become an engineer ticket.
- already landed - equivalent behavior/docs are on main through another commit.
- obsolete/superseded - the work is no longer desired because the design moved on.
- unclear - needs Kai decision.
Post one comment on this issue with a concise per-branch table-free report: branch, likely issue, classification, evidence, and recommended next issue title if recovery is needed.

Do not land code from this advisor pass. File no recovery issues unless the classification is obvious and low-risk; otherwise leave recommended titles.

Filed from the read-only director surface (she/her).

## Goal Audit the old `ward-salvage/*` branches that still contain real branch-only changes and determine whether each holds recoverable lost work, has already landed through another path, or is obsolete/superseded. ## Branches to inspect * `ward-salvage/ward-0d7db064` - branch-only files suggest error reporting work. * `ward-salvage/ward-33ebc07c` - workflow mirror files/docs. * `ward-salvage/ward-71d4c8f9` - agent reap implementation files. * `ward-salvage/ward-86ffb40e` - agent adapter/fleet test changes. * `ward-salvage/ward-9b76e536` - director consult files/docs. * `ward-salvage/ward-df8343fa` - graft-removal doc differs from main. * `ward-salvage/ward-e6ca4924` - advisor/reply/doc changes. ## Context A director surface audit found five other salvage branches are provenance-only and tied to already closed issues: `ward-11906072`, `ward-9686d657`, `ward-9c289e1b`, `ward-d9e51395`, and `ward-7ca66e36`. Do not spend time on those unless needed as comparison. The seven branches above are concerning because their branch-only diffs include real code/doc changes from about a week ago. Some may correspond to closed issues whose work later landed differently. Others may be lost work that should be recovered. ## Do For each branch: 1. Fetch the branch and compare it to current `origin/main`. 2. Identify the likely source issue from commit messages, file names, and issue search. 3. Decide one of: * `recover` - useful work is not on main and should become an engineer ticket. * `already landed` - equivalent behavior/docs are on main through another commit. * `obsolete/superseded` - the work is no longer desired because the design moved on. * `unclear` - needs Kai decision. 4. Post one comment on this issue with a concise per-branch table-free report: branch, likely issue, classification, evidence, and recommended next issue title if recovery is needed. Do not land code from this advisor pass. File no recovery issues unless the classification is obvious and low-risk; otherwise leave recommended titles. Filed from the read-only director surface (she/her).

coilyco-ops added the

consult

headless

labels

2026-07-08 08:39:37 +00:00

coilyco-ops commented

2026-07-08 08:44:23 +00:00

Author

Member

🔎 ward agent advisor

ward agent advisor --harness codex ran a one-shot standard research pass on this question:

Audit the seven ward-salvage branches listed in the issue. Fetch and compare each branch to current origin/main, identify the likely source issue, classify as recover, already landed, obsolete/superseded, or unclear, and post one concise branch-by-branch comment with evidence and recommended recovery issue titles where needed. Do not land code.

ward-salvage/ward-0d7db064
Likely source: ward#303 - ward => bugsnag / sentry integration.
Classification: unclear.
Evidence: the branch adds a real crash-reporting implementation (cmd/ward/errorreport.go, docs/error-reporting.md, getsentry/sentry-go in go.mod), and none of that exists on origin/main. But the live ward#303 issue already contains a NO-GO note saying the issue was too underspecified to choose a provider or scope. This salvage picks Sentry / GlitchTip and a panic-only boundary without that decision having landed on main, so the code is recoverable only after Kai decides the product choice.
Recommended next issue title: Decide ward crash-reporting provider and scope, then recover salvage error-reporting work.
ward-salvage/ward-33ebc07c
Likely source: ward#214 - Mirror forgejo actions into GitHub actions.
Classification: recover.
Evidence: current main still has only .forgejo/workflows/release.yml on Forgejo and only .github/workflows/test.yml on GitHub. The salvage branch adds the missing .forgejo/workflows/test.yml, a drift checker (scripts/check_workflow_mirror.py), a Go test, a pre-commit hook, and docs. The live ward#214 research comment says the useful direction is actually GitHub -> Forgejo mirroring of test.yml, because canonical Forgejo pushes are otherwise unvetted. The branch is stale in places and should not be cherry-picked blind, but it clearly contains recoverable work.
Recommended next issue title: Mirror the GitHub test workflow onto Forgejo and add a workflow-drift linter.
ward-salvage/ward-71d4c8f9
Likely source: ward#376 - ward agent: reap engineer containers idle >1h.
Classification: already landed.
Evidence: origin/main already contains cmd/ward/agent_reap.go, cmd/ward/agent_reap_compute.go, cmd/ward/agent_reap_test.go, and docs/agent-reap.md. The feature landed as 559d6ed (feat(agent): add \ward agent reap` host-side idle-killer for wedged engineers) and was later refined by follow-up reap fixes (231cbb6, b9b1885, 703ebe3, 343abb5). The salvage branch is an older residual version of functionality that is now on main` in a more evolved form.
ward-salvage/ward-86ffb40e
Likely source: the closed cleanup chain around ward#206, ward#420, ward#419, and ward#423.
Classification: obsolete/superseded.
Evidence: this branch deletes agent_adapter.go and transitional registry / signature switches, and adds a new fleet_test.go golden-fixture test. Current main already landed the parseMode part of ward#420 - parseMode now validates against fleet.agents - and later landed a different fleet-test cleanup as 98ead4e (test(fleet): replace hand-synced golden fixture with structural invariants, closes #423). At the same time, current docs and code still intentionally keep agent_adapter.go and the manifest projection alive. So this salvage is not missing product work. It is an abandoned intermediate cleanup path that current main has partly replaced and partly rejected.
ward-salvage/ward-9b76e536
Likely source: ward#493 - warded director: consult-to-headless conversion flow.
Classification: already landed.
Evidence: current main already has cmd/ward/agent_director_consult.go, its tests, and docs/director-consult.md. The feature landed as 948a576 (feat(director): consult-to-headless conversion interview) and later got follow-up repair in a413e70 (fix(agent): allow consult dispatch). The salvage branch is an earlier residual implementation of the same feature.
ward-salvage/ward-df8343fa
Likely source: ward#407 - Remove the ward-over-ward-kdl graft layer.
Classification: already landed.
Evidence: the only branch diff is docs/ops-forgejo-graft-removal.md, and current main already carries that doc via 88b445c (docs(ops): answer the ward#407 graft-removal consult + add the removal guardrail). The salvage copy is an older draft. The live doc on main is the landed version.
ward-salvage/ward-e6ca4924
Likely source: ward#411 - advisor ref mode runs native (host one-shot); containerize it like engineer + freeform modes.
Classification: already landed.
Evidence: current main already describes and implements advisor ref-mode research in a fresh container. The main landing is 6cbed47 (feat(advisor): containerize ref-mode research, closes #411), and 48a3375 later fixed broker-safe capture behavior. The salvage branch matches that theme exactly (captureAgentContainer, reply-plan containerization, docs updates), but main already has the feature and later fixes.

Overall: 2 branches look worth reopening as fresh work from current main (ward-33ebc07c recover, ward-0d7db064 unclear pending a product decision). The other 5 are not lost work that should be recovered as-is: 4 already landed through later commits, and ward-86ffb40e is an outdated cleanup path superseded by later fleet / registry work.

Researched and posted automatically by ward agent advisor --harness codex (ward#179). This is one-shot research, not a carried change - verify before acting on it.

— Codex, via ward agent

### 🔎 ward agent advisor `ward agent advisor --harness codex` ran a one-shot **standard** research pass on this question: > Audit the seven ward-salvage branches listed in the issue. Fetch and compare each branch to current origin/main, identify the likely source issue, classify as recover, already landed, obsolete/superseded, or unclear, and post one concise branch-by-branch comment with evidence and recommended recovery issue titles where needed. Do not land code. --- - `ward-salvage/ward-0d7db064` Likely source: `ward#303` - `ward => bugsnag / sentry integration`. Classification: `unclear`. Evidence: the branch adds a real crash-reporting implementation (`cmd/ward/errorreport.go`, `docs/error-reporting.md`, `getsentry/sentry-go` in `go.mod`), and none of that exists on `origin/main`. But the live `ward#303` issue already contains a NO-GO note saying the issue was too underspecified to choose a provider or scope. This salvage picks `Sentry` / `GlitchTip` and a panic-only boundary without that decision having landed on `main`, so the code is recoverable only after Kai decides the product choice. Recommended next issue title: `Decide ward crash-reporting provider and scope, then recover salvage error-reporting work`. - `ward-salvage/ward-33ebc07c` Likely source: `ward#214` - `Mirror forgejo actions into GitHub actions`. Classification: `recover`. Evidence: current `main` still has only `.forgejo/workflows/release.yml` on Forgejo and only `.github/workflows/test.yml` on GitHub. The salvage branch adds the missing `.forgejo/workflows/test.yml`, a drift checker (`scripts/check_workflow_mirror.py`), a Go test, a pre-commit hook, and docs. The live `ward#214` research comment says the useful direction is actually GitHub -> Forgejo mirroring of `test.yml`, because canonical Forgejo pushes are otherwise unvetted. The branch is stale in places and should not be cherry-picked blind, but it clearly contains recoverable work. Recommended next issue title: `Mirror the GitHub test workflow onto Forgejo and add a workflow-drift linter`. - `ward-salvage/ward-71d4c8f9` Likely source: `ward#376` - `ward agent: reap engineer containers idle >1h`. Classification: `already landed`. Evidence: `origin/main` already contains `cmd/ward/agent_reap.go`, `cmd/ward/agent_reap_compute.go`, `cmd/ward/agent_reap_test.go`, and `docs/agent-reap.md`. The feature landed as `559d6ed` (`feat(agent): add \`ward agent reap\` host-side idle-killer for wedged engineers`) and was later refined by follow-up reap fixes (`231cbb6`, `b9b1885`, `703ebe3`, `343abb5`). The salvage branch is an older residual version of functionality that is now on `main` in a more evolved form. - `ward-salvage/ward-86ffb40e` Likely source: the closed cleanup chain around `ward#206`, `ward#420`, `ward#419`, and `ward#423`. Classification: `obsolete/superseded`. Evidence: this branch deletes `agent_adapter.go` and transitional registry / signature switches, and adds a new `fleet_test.go` golden-fixture test. Current `main` already landed the `parseMode` part of `ward#420` - `parseMode` now validates against `fleet.agents` - and later landed a different fleet-test cleanup as `98ead4e` (`test(fleet): replace hand-synced golden fixture with structural invariants`, closes `#423`). At the same time, current docs and code still intentionally keep `agent_adapter.go` and the manifest projection alive. So this salvage is not missing product work. It is an abandoned intermediate cleanup path that current `main` has partly replaced and partly rejected. - `ward-salvage/ward-9b76e536` Likely source: `ward#493` - `warded director: consult-to-headless conversion flow`. Classification: `already landed`. Evidence: current `main` already has `cmd/ward/agent_director_consult.go`, its tests, and `docs/director-consult.md`. The feature landed as `948a576` (`feat(director): consult-to-headless conversion interview`) and later got follow-up repair in `a413e70` (`fix(agent): allow consult dispatch`). The salvage branch is an earlier residual implementation of the same feature. - `ward-salvage/ward-df8343fa` Likely source: `ward#407` - `Remove the ward-over-ward-kdl graft layer`. Classification: `already landed`. Evidence: the only branch diff is `docs/ops-forgejo-graft-removal.md`, and current `main` already carries that doc via `88b445c` (`docs(ops): answer the ward#407 graft-removal consult + add the removal guardrail`). The salvage copy is an older draft. The live doc on `main` is the landed version. - `ward-salvage/ward-e6ca4924` Likely source: `ward#411` - `advisor ref mode runs native (host one-shot); containerize it like engineer + freeform modes`. Classification: `already landed`. Evidence: current `main` already describes and implements advisor ref-mode research in a fresh container. The main landing is `6cbed47` (`feat(advisor): containerize ref-mode research`, closes `#411`), and `48a3375` later fixed broker-safe capture behavior. The salvage branch matches that theme exactly (`captureAgentContainer`, reply-plan containerization, docs updates), but `main` already has the feature and later fixes. Overall: 2 branches look worth reopening as fresh work from current `main` (`ward-33ebc07c` recover, `ward-0d7db064` unclear pending a product decision). The other 5 are not lost work that should be recovered as-is: 4 already landed through later commits, and `ward-86ffb40e` is an outdated cleanup path superseded by later fleet / registry work. --- Researched and posted automatically by `ward agent advisor --harness codex` (ward#179). This is one-shot research, not a carried change - verify before acting on it.   — Codex, via `ward agent`

coilyco-ops removed the

consult

label

2026-07-08 08:44:35 +00:00

coilyco-ops commented

2026-07-08 08:44:43 +00:00

Author

Member

🔒 Reserved by ward agent --harness codex — container engineer-codex-ward-666 on host kais-macbook-pro-2.local is carrying this issue (reserved 2026-07-08T08:44:41Z). Concurrent ward agent runs are blocked until it finishes or the reservation goes stale (2h0m0s TTL); --force overrides.

Do not comment on or edit this issue to steer the run while it is reserved. The engineer seeded the body once at launch and never re-reads it, so a comment or edit reaches only human readers, never the running engineer. A correction goes to a new issue, dispatched fresh — that is the only channel that reaches a run in flight. Where the forge supports it, ward locks this conversation to make that a road-block rather than a convention (ward#494).

run seed context — what this run is carrying (ward#609)

Resolved: coilyco-flight-deck/ward#666 · branch issue-666 · harness codex · workflow direct-main
Run: engineer-codex-ward-666 · ward v0.443.0 · dispatched 2026-07-08T08:44:41Z
Comment thread: 1 included in the pre-flight read, 0 stripped (ward's own automated comments).
- included: @coilyco-ops (2026-07-08T08:44:23Z)

Issue body as seeded:

## Goal

Audit the old `ward-salvage/*` branches that still contain real branch-only changes and determine whether each holds recoverable lost work, has already landed through another path, or is obsolete/superseded.

## Branches to inspect

* `ward-salvage/ward-0d7db064` - branch-only files suggest error reporting work.
* `ward-salvage/ward-33ebc07c` - workflow mirror files/docs.
* `ward-salvage/ward-71d4c8f9` - agent reap implementation files.
* `ward-salvage/ward-86ffb40e` - agent adapter/fleet test changes.
* `ward-salvage/ward-9b76e536` - director consult files/docs.
* `ward-salvage/ward-df8343fa` - graft-removal doc differs from main.
* `ward-salvage/ward-e6ca4924` - advisor/reply/doc changes.

## Context

A director surface audit found five other salvage branches are provenance-only and tied to already closed issues: `ward-11906072`, `ward-9686d657`, `ward-9c289e1b`, `ward-d9e51395`, and `ward-7ca66e36`. Do not spend time on those unless needed as comparison.

The seven branches above are concerning because their branch-only diffs include real code/doc changes from about a week ago. Some may correspond to closed issues whose work later landed differently. Others may be lost work that should be recovered.

## Do

For each branch:

1. Fetch the branch and compare it to current `origin/main`.
2. Identify the likely source issue from commit messages, file names, and issue search.
3. Decide one of:
   * `recover` - useful work is not on main and should become an engineer ticket.
   * `already landed` - equivalent behavior/docs are on main through another commit.
   * `obsolete/superseded` - the work is no longer desired because the design moved on.
   * `unclear` - needs Kai decision.
4. Post one comment on this issue with a concise per-branch table-free report: branch, likely issue, classification, evidence, and recommended next issue title if recovery is needed.

Do not land code from this advisor pass. File no recovery issues unless the classification is obviou

… (truncated to 2000 chars; full body is on this issue)

Static container doctrine and seed boilerplate are identical every run and omitted here (they ride ward v0.443.0).

— Codex, via ward agent

🔒 Reserved by `ward agent --harness codex` — container `engineer-codex-ward-666` on host `kais-macbook-pro-2.local` is carrying this issue (reserved 2026-07-08T08:44:41Z). Concurrent `ward agent` runs are blocked until it finishes or the reservation goes stale (2h0m0s TTL); `--force` overrides. **Do not comment on or edit this issue to steer the run while it is reserved.** The engineer seeded the body once at launch and never re-reads it, so a comment or edit reaches only human readers, never the running engineer. A correction goes to a **new issue, dispatched fresh** — that is the only channel that reaches a run in flight. Where the forge supports it, ward locks this conversation to make that a road-block rather than a convention (ward#494). <details><summary>run seed context — what this run is carrying (ward#609)</summary> - **Resolved:** `coilyco-flight-deck/ward#666` · branch `issue-666` · harness `codex` · workflow `direct-main` - **Run:** `engineer-codex-ward-666` · ward `v0.443.0` · dispatched `2026-07-08T08:44:41Z` - **Comment thread:** 1 included in the pre-flight read, 0 stripped (ward's own automated comments). - included: @coilyco-ops (2026-07-08T08:44:23Z) **Issue body as seeded:** ``` ## Goal Audit the old `ward-salvage/*` branches that still contain real branch-only changes and determine whether each holds recoverable lost work, has already landed through another path, or is obsolete/superseded. ## Branches to inspect * `ward-salvage/ward-0d7db064` - branch-only files suggest error reporting work. * `ward-salvage/ward-33ebc07c` - workflow mirror files/docs. * `ward-salvage/ward-71d4c8f9` - agent reap implementation files. * `ward-salvage/ward-86ffb40e` - agent adapter/fleet test changes. * `ward-salvage/ward-9b76e536` - director consult files/docs. * `ward-salvage/ward-df8343fa` - graft-removal doc differs from main. * `ward-salvage/ward-e6ca4924` - advisor/reply/doc changes. ## Context A director surface audit found five other salvage branches are provenance-only and tied to already closed issues: `ward-11906072`, `ward-9686d657`, `ward-9c289e1b`, `ward-d9e51395`, and `ward-7ca66e36`. Do not spend time on those unless needed as comparison. The seven branches above are concerning because their branch-only diffs include real code/doc changes from about a week ago. Some may correspond to closed issues whose work later landed differently. Others may be lost work that should be recovered. ## Do For each branch: 1. Fetch the branch and compare it to current `origin/main`. 2. Identify the likely source issue from commit messages, file names, and issue search. 3. Decide one of: * `recover` - useful work is not on main and should become an engineer ticket. * `already landed` - equivalent behavior/docs are on main through another commit. * `obsolete/superseded` - the work is no longer desired because the design moved on. * `unclear` - needs Kai decision. 4. Post one comment on this issue with a concise per-branch table-free report: branch, likely issue, classification, evidence, and recommended next issue title if recovery is needed. Do not land code from this advisor pass. File no recovery issues unless the classification is obviou ``` … (truncated to 2000 chars; full body is on this issue) Static container doctrine and seed boilerplate are identical every run and omitted here (they ride ward v0.443.0). </details>  — Codex, via `ward agent`

coilyco-ops referenced this issue

2026-07-08 08:45:20 +00:00

Recover workflow mirror work from ward-salvage/ward-33ebc07c #667

coilyco-ops referenced this issue

2026-07-08 08:45:30 +00:00

Decide ward crash-reporting provider and scope before recovering salvage error reporting #668

coilyco-ops closed this issue

2026-07-08 08:48:12 +00:00

coilyco-ops referenced this issue from a commit

2026-07-08 08:52:35 +00:00

feat(ci): recover workflow mirror

coilyco-ops referenced this issue from a commit