Split issueForge into a git-host port and a Tracker port

coilyco-ops commented

2026-07-07 21:46:53 +00:00

Owner

Parent epic: #633. Depends on: none. This unblocks the Trello (#636) and Shortcut (#637) tracker adapters and the Jira/Linear backlog (#638).

Context. cmd/ward/forge.go defines issueForge and a forge enum (forgeForgejo default, forgeGitHub). That single port bundles two responsibilities that are one system on Forgejo and GitHub but two systems on most other stacks: the git host (clone origin, push credentials, PR/MR) and the issue tracker (list, read, comment, close, triage). A GitHub-repo-plus-Trello-board user needs the git host to be GitHub and the issue thread to be Trello. Today that pairing is unexpressible without editing core flow.

Proposed change.

Extract a git-host port (call it gitHost or keep forge for it): clone base URL, push user, ref parsing, PR/MR creation. Forgejo and GitHub keep their current implementations here.
Extract a Tracker port: read an issue, post a comment, close, apply triage/labels, resolve an issue ref to a stable URL. Forgejo issues and GitHub issues become the first two Tracker implementations, so nothing regresses.
A run pairs one git host with one tracker. On Forgejo and GitHub the pair is the same system, which is the default and must stay zero-config.
Route the existing issueForge call sites (hostForgeClient, the agent_* files that read and comment on issues) through whichever half they actually need.

Acceptance criteria.

Forgejo-only and GitHub-only runs behave exactly as today, defaults unchanged, WARD_FORGE still selects the git host.
A test proves a git host and a tracker can be paired independently (even if the second tracker is a stub in this issue).
No agent_* file talks to a concrete forge directly. All issue and comment traffic goes through the Tracker port, all clone/push/PR traffic through the git-host port.
docs/compat-surface.md (#640) updated to show the two ports.

Notes. Shape the Tracker interface against a real second tracker (Trello, #636) so it is not designed in the abstract. Keep the method set minimal: what an agent run genuinely does to an issue, nothing more.

Milestone: foreign-stack compat. Filed from the read-only director surface (she/her), part of the release-readiness issue batch.

Parent epic: #633. **Depends on:** none. This unblocks the Trello (#636) and Shortcut (#637) tracker adapters and the Jira/Linear backlog (#638). **Context.** `cmd/ward/forge.go` defines `issueForge` and a `forge` enum (`forgeForgejo` default, `forgeGitHub`). That single port bundles two responsibilities that are one system on Forgejo and GitHub but two systems on most other stacks: the git host (clone origin, push credentials, PR/MR) and the issue tracker (list, read, comment, close, triage). A GitHub-repo-plus-Trello-board user needs the git host to be GitHub and the issue thread to be Trello. Today that pairing is unexpressible without editing core flow. **Proposed change.** - Extract a git-host port (call it `gitHost` or keep `forge` for it): clone base URL, push user, ref parsing, PR/MR creation. Forgejo and GitHub keep their current implementations here. - Extract a `Tracker` port: read an issue, post a comment, close, apply triage/labels, resolve an issue ref to a stable URL. Forgejo issues and GitHub issues become the first two `Tracker` implementations, so nothing regresses. - A run pairs one git host with one tracker. On Forgejo and GitHub the pair is the same system, which is the default and must stay zero-config. - Route the existing `issueForge` call sites (`hostForgeClient`, the `agent_*` files that read and comment on issues) through whichever half they actually need. **Acceptance criteria.** - Forgejo-only and GitHub-only runs behave exactly as today, defaults unchanged, `WARD_FORGE` still selects the git host. - A test proves a git host and a tracker can be paired independently (even if the second tracker is a stub in this issue). - No `agent_*` file talks to a concrete forge directly. All issue and comment traffic goes through the `Tracker` port, all clone/push/PR traffic through the git-host port. - `docs/compat-surface.md` (#640) updated to show the two ports. **Notes.** Shape the `Tracker` interface against a real second tracker (Trello, #636) so it is not designed in the abstract. Keep the method set minimal: what an agent run genuinely does to an issue, nothing more. --- Milestone: foreign-stack compat. Filed from the read-only director surface (she/her), part of the release-readiness issue batch.

coilyco-ops added this to the foreign-stack compat milestone

2026-07-07 21:46:53 +00:00

coilyco-ops added the

labels

2026-07-07 21:46:53 +00:00

coilyco-ops referenced this issue

2026-07-07 21:47:16 +00:00

GitLab forge adapter #635

coilyco-ops referenced this issue

2026-07-07 21:47:16 +00:00

Trello tracker adapter #636

coilyco-ops referenced this issue

2026-07-07 21:47:16 +00:00

Shortcut tracker adapter #637

coilyco-ops referenced this issue

2026-07-07 21:47:16 +00:00

Additional tracker adapters (Jira, Linear) #638

coilyco-ops referenced this issue

2026-07-07 21:47:17 +00:00

Add docs/compat-surface.md and make the compat surface first-class in CONTRIBUTING #640

coilyco-ops referenced this issue

2026-07-07 21:48:40 +00:00

[epic] Release readiness = running ward on foreign stacks (split-stack: git host + tracker) #633

coilyco-ops added

and removed

labels

2026-07-09 18:25:07 +00:00

coilyco-ops commented

2026-07-09 18:25:08 +00:00

Author

Owner

Consult interrogation: this is now headless-ready. The issue has a concrete implementation target and acceptance criteria, with no remaining human decision needed before an engineer can carry it.

coilyco-ops referenced this issue

2026-07-09 18:26:59 +00:00

Choose an out-of-band issue tracker source before public release #827

coilyco-ops commented

2026-07-09 18:29:53 +00:00

Author

Owner

WARD-RESERVATION: held 🔒

reservation details

Holder: container engineer-codex-ward-634 on host kais-macbook-pro-2.local.

Reserved by ward agent --harness codex (reserved 2026-07-09T18:29:51Z). Concurrent ward agent runs are blocked until it finishes or the reservation goes stale (1h 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#634 · branch issue-634 · harness codex · workflow pull-request
Run: engineer-codex-ward-634 · ward v0.493.0 · dispatched 2026-07-09T18:29:51Z
Comment thread: 1 included in the pre-flight read, 0 stripped (ward's own automated comments).
- included: @coilyco-ops (2026-07-09T18:25:08Z)

Static container doctrine and seed boilerplate are identical every run and omitted here (they ride ward v0.493.0).

— Codex, via ward agent

WARD-RESERVATION: held 🔒 <details><summary>reservation details</summary> Holder: container `engineer-codex-ward-634` on host `kais-macbook-pro-2.local`. Reserved by `ward agent --harness codex` (reserved 2026-07-09T18:29:51Z). Concurrent `ward agent` runs are blocked until it finishes or the reservation goes stale (1h 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#634` · branch `issue-634` · harness `codex` · workflow `pull-request` - **Run:** `engineer-codex-ward-634` · ward `v0.493.0` · dispatched `2026-07-09T18:29:51Z` - **Comment thread:** 1 included in the pre-flight read, 0 stripped (ward's own automated comments). - included: @coilyco-ops (2026-07-09T18:25:08Z) Static container doctrine and seed boilerplate are identical every run and omitted here (they ride ward v0.493.0). </details> </details>  — Codex, via `ward agent`

coilyco-ops referenced this issue from a pull request that will close it,

2026-07-09 18:49:12 +00:00

Split issueForge into a git-host port and a Tracker port #856

coilyco-ops commented

2026-07-09 18:53:59 +00:00

Author

Owner

WARD-OUTCOME: done ✅

details

workflow: pull-request; review summary: review gate skipped by ~/.ward/config.yaml default felt: straightforward split, one lint pass needed cleanup confidence: high surprise: golangci-lint caught the legacy helper and cyclo edge follow-ups: none

WARD-OUTCOME: done ✅ <details><summary>details</summary> workflow: pull-request; review summary: review gate skipped by ~/.ward/config.yaml default felt: straightforward split, one lint pass needed cleanup confidence: high surprise: golangci-lint caught the legacy helper and cyclo edge follow-ups: none </details>

coilyco-ops commented

2026-07-09 20:12:02 +00:00

Author

Owner

WARD-RESERVATION: held 🔒

reservation details

Holder: container engineer-codex-ward-634 on host kais-macbook-pro-2.local.

Reserved by ward agent --harness codex (reserved 2026-07-09T20:12:00Z). Concurrent ward agent runs are blocked until it finishes or the reservation goes stale (1h 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#634 · branch issue-634 · harness codex · workflow direct-to-main
Run: engineer-codex-ward-634 · ward v0.522.0 · dispatched 2026-07-09T20:12:00Z
Comment thread: 2 included in the pre-flight read, 1 stripped (ward's own automated comments).
- included: @coilyco-ops (2026-07-09T18:25:08Z), @coilyco-ops (2026-07-09T18:53:59Z)
- stripped: @coilyco-ops (2026-07-09T18:29:53Z)

Static container doctrine and seed boilerplate are identical every run and omitted here (they ride ward v0.522.0).

— Codex, via ward agent

WARD-RESERVATION: held 🔒 <details><summary>reservation details</summary> Holder: container `engineer-codex-ward-634` on host `kais-macbook-pro-2.local`. Reserved by `ward agent --harness codex` (reserved 2026-07-09T20:12:00Z). Concurrent `ward agent` runs are blocked until it finishes or the reservation goes stale (1h 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#634` · branch `issue-634` · harness `codex` · workflow `direct-to-main` - **Run:** `engineer-codex-ward-634` · ward `v0.522.0` · dispatched `2026-07-09T20:12:00Z` - **Comment thread:** 2 included in the pre-flight read, 1 stripped (ward's own automated comments). - included: @coilyco-ops (2026-07-09T18:25:08Z), @coilyco-ops (2026-07-09T18:53:59Z) - stripped: @coilyco-ops (2026-07-09T18:29:53Z) Static container doctrine and seed boilerplate are identical every run and omitted here (they ride ward v0.522.0). </details> </details>  — Codex, via `ward agent`

coilyco-ops referenced this issue from a commit

2026-07-09 20:20:44 +00:00

closes #634

coilyco-ops closed this issue

2026-07-09 20:20:44 +00:00

coilyco-ops commented

2026-07-09 20:21:08 +00:00

Author

Owner

WARD-OUTCOME: done

details

workflow: direct-to-main; review summary: skipped because review gate skipped by ~/.ward/config.yaml default

Implemented the git-host and tracker split, kept Forgejo/GitHub paired by default, and landed it on main.

The implementation felt mostly mechanical after the merge conflict was resolved. The only surprise was origin/main moving while I was landing, so I merged that tip before pushing.

Confidence: high.
Follow-ups: none.

WARD-OUTCOME: done <details><summary>details</summary> workflow: direct-to-main; review summary: skipped because review gate skipped by ~/.ward/config.yaml default Implemented the git-host and tracker split, kept Forgejo/GitHub paired by default, and landed it on main. The implementation felt mostly mechanical after the merge conflict was resolved. The only surprise was origin/main moving while I was landing, so I merged that tip before pushing. Confidence: high. Follow-ups: none. </details>

coilyco-ops referenced this issue

2026-07-09 20:27:37 +00:00

Split issueForge into a git-host port and a Tracker port #856

coilysiren referenced this issue

2026-07-14 15:27:09 +00:00

director should emit colorized, space buffered, visually distinct messages about the categories of actions its performing on startup #1283

coilyco-ops referenced this issue

2026-07-23 04:53:12 +00:00

Make split-stack engineer dispatch preserve tracker authority and verify git-host drift #1519

Rows
Columns

Split issueForge into a git-host port and a Tracker port #634