From 77a918aa9912e54f94a61876c55c6f8a90cf201e Mon Sep 17 00:00:00 2001 From: Luke Curley Date: Sat, 26 Sep 2026 11:28:54 -0700 Subject: [PATCH 01/11] skill(spawn-quests): surface open decisions as prompts, keep PRs draft until resolved Co-Authored-By: Claude Opus 5.5 --- .claude/skills/spawn-quests/SKILL.md | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) diff --git a/.claude/skills/spawn-quests/SKILL.md b/.claude/skills/spawn-quests/SKILL.md index b8e0f44ffab..1c036398d47 100644 --- a/.claude/skills/spawn-quests/SKILL.md +++ b/.claude/skills/spawn-quests/SKILL.md @@ -26,8 +26,12 @@ Each agent blocks on its own checks and reports back only when done or blocked. Limit the concurrency to at most N agents in parallel, where N is half the number of physical CPU cores. Other sessions share this machine: hold new agents while the load average exceeds the core count. +Each agent keeps its PR a draft while any decision is open, and lists those decisions in its report. Report each sub-agent's final status, staying silent on interim notifications, but do not monitor their PRs. -Prompt the user if they want to /plan-quests for any suggested follow-ups. + +At the end, surface every open decision as an interactive prompt: opinions (naming, API shape, branch), blockers (releases, approvals, manual steps), and follow-ups to /plan-quests. +Batch a few per prompt, each with the PR, a short summary, and your recommendation. +Relay each answer to the agent that owns the PR; it marks the PR ready only once all of its decisions are resolved. Run /plan-quests for any selected quests in the foreground. Perform any research and monitoring in the background. From bc6db0d6b5fce19da49019d6a15c52785e9c9b45 Mon Sep 17 00:00:00 2001 From: Luke Curley Date: Sat, 26 Sep 2026 11:33:01 -0700 Subject: [PATCH 02/11] skill(spawn-quests): prompt as each agent finishes Co-Authored-By: Claude Opus 5.5 --- .claude/skills/spawn-quests/SKILL.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/.claude/skills/spawn-quests/SKILL.md b/.claude/skills/spawn-quests/SKILL.md index 1c036398d47..d9a0ca687c6 100644 --- a/.claude/skills/spawn-quests/SKILL.md +++ b/.claude/skills/spawn-quests/SKILL.md @@ -29,7 +29,7 @@ Other sessions share this machine: hold new agents while the load average exceed Each agent keeps its PR a draft while any decision is open, and lists those decisions in its report. Report each sub-agent's final status, staying silent on interim notifications, but do not monitor their PRs. -At the end, surface every open decision as an interactive prompt: opinions (naming, API shape, branch), blockers (releases, approvals, manual steps), and follow-ups to /plan-quests. +When an agent finishes, surface its open decisions as interactive prompts: opinions (naming, API shape, branch), blockers (releases, approvals, manual steps), and follow-ups to /plan-quests. Batch a few per prompt, each with the PR, a short summary, and your recommendation. Relay each answer to the agent that owns the PR; it marks the PR ready only once all of its decisions are resolved. From a9e3cc75dd0f0109fadb853fb6baa21dd741286c Mon Sep 17 00:00:00 2001 From: Luke Curley Date: Sat, 26 Sep 2026 11:42:11 -0700 Subject: [PATCH 03/11] skill(spawn-quests): prompt inline per report, coordinator marks PRs ready Co-Authored-By: Claude Opus 5.5 --- .claude/skills/spawn-quests/SKILL.md | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) diff --git a/.claude/skills/spawn-quests/SKILL.md b/.claude/skills/spawn-quests/SKILL.md index d9a0ca687c6..21e8e6bd1ea 100644 --- a/.claude/skills/spawn-quests/SKILL.md +++ b/.claude/skills/spawn-quests/SKILL.md @@ -26,12 +26,13 @@ Each agent blocks on its own checks and reports back only when done or blocked. Limit the concurrency to at most N agents in parallel, where N is half the number of physical CPU cores. Other sessions share this machine: hold new agents while the load average exceeds the core count. -Each agent keeps its PR a draft while any decision is open, and lists those decisions in its report. +Each agent opens its PR as a draft, lists every open decision in its report with a recommendation, and never marks it ready itself. Report each sub-agent's final status, staying silent on interim notifications, but do not monitor their PRs. -When an agent finishes, surface its open decisions as interactive prompts: opinions (naming, API shape, branch), blockers (releases, approvals, manual steps), and follow-ups to /plan-quests. +As each agent reports, prompt the user inline, without waiting for the rest: opinions (naming, API shape, branch), blockers (releases, approvals, manual steps), and follow-ups to /plan-quests. Batch a few per prompt, each with the PR, a short summary, and your recommendation. -Relay each answer to the agent that owns the PR; it marks the PR ready only once all of its decisions are resolved. +Resume the owning agent with each answer, and prompt again on anything its next report raises. +Mark the PR ready yourself once all of its decisions are resolved. Run /plan-quests for any selected quests in the foreground. Perform any research and monitoring in the background. From a8dcac8c187d7c3fb59810408c050bba4bbf2d88 Mon Sep 17 00:00:00 2001 From: Luke Curley Date: Sat, 26 Sep 2026 11:44:46 -0700 Subject: [PATCH 04/11] skill(start-quest): keep PRs draft until the user confirms every decision Co-Authored-By: Claude Opus 5.5 --- .claude/skills/start-quest/SKILL.md | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/.claude/skills/start-quest/SKILL.md b/.claude/skills/start-quest/SKILL.md index 73d0911e40b..c3ff5844946 100644 --- a/.claude/skills/start-quest/SKILL.md +++ b/.claude/skills/start-quest/SKILL.md @@ -18,5 +18,5 @@ Keep scratch files (PR body, logs, notes) in the worktree's gitignored `.scratch Never write to or clean up a directory other agents share, such as a session scratchpad. When done, summarize any issues encounted, and suggest potential follow-up. -If you're happy with the outcome, switch the draft PR to ready for review. -If you want another set of eyes on it, keep it a draft. +List every open decision (naming, API shape, branch, blockers, manual steps) with a recommendation, and prompt the user interactively when you can. +Keep the PR a draft until the user has confirmed every decision, then mark it ready for review. From 9a133498367ae5f4fc8b022b42be7ac423f825ff Mon Sep 17 00:00:00 2001 From: Luke Curley Date: Sat, 26 Sep 2026 13:56:12 -0700 Subject: [PATCH 05/11] skill(spawn-quests,start-quest): explain each result, prompt decisions, ready, and follow-ups Co-Authored-By: Claude Opus 5.5 --- .claude/skills/spawn-quests/SKILL.md | 5 +++-- .claude/skills/start-quest/SKILL.md | 4 ++-- 2 files changed, 5 insertions(+), 4 deletions(-) diff --git a/.claude/skills/spawn-quests/SKILL.md b/.claude/skills/spawn-quests/SKILL.md index 21e8e6bd1ea..cfcf91e6d4b 100644 --- a/.claude/skills/spawn-quests/SKILL.md +++ b/.claude/skills/spawn-quests/SKILL.md @@ -29,8 +29,9 @@ Other sessions share this machine: hold new agents while the load average exceed Each agent opens its PR as a draft, lists every open decision in its report with a recommendation, and never marks it ready itself. Report each sub-agent's final status, staying silent on interim notifications, but do not monitor their PRs. -As each agent reports, prompt the user inline, without waiting for the rest: opinions (naming, API shape, branch), blockers (releases, approvals, manual steps), and follow-ups to /plan-quests. -Batch a few per prompt, each with the PR, a short summary, and your recommendation. +As each agent reports, explain its result in a few lines, then prompt the user inline without waiting for the rest. +Ask about its decisions (naming, API shape, branch, blockers, manual steps), each with the PR, a short summary, and your recommendation, then offer its follow-ups as a multi-select to /plan-quests. +When nothing is open, ask whether to mark the PR ready. Resume the owning agent with each answer, and prompt again on anything its next report raises. Mark the PR ready yourself once all of its decisions are resolved. diff --git a/.claude/skills/start-quest/SKILL.md b/.claude/skills/start-quest/SKILL.md index c3ff5844946..9b9742fc1a5 100644 --- a/.claude/skills/start-quest/SKILL.md +++ b/.claude/skills/start-quest/SKILL.md @@ -17,6 +17,6 @@ Implement the quest until it is complete, or some blocker is hit, then create a Keep scratch files (PR body, logs, notes) in the worktree's gitignored `.scratch/`. Never write to or clean up a directory other agents share, such as a session scratchpad. -When done, summarize any issues encounted, and suggest potential follow-up. -List every open decision (naming, API shape, branch, blockers, manual steps) with a recommendation, and prompt the user interactively when you can. +When done, explain the result in a few lines and summarize any issues encountered. +List every open decision (naming, API shape, branch, blockers, manual steps) with a recommendation, and prompt the user interactively when you can, offering follow-ups as a multi-select to /plan-quests. Keep the PR a draft until the user has confirmed every decision, then mark it ready for review. From fd5b4ae3e80051f8f59109ddd256bccddc2386d3 Mon Sep 17 00:00:00 2001 From: Luke Curley Date: Sat, 26 Sep 2026 13:59:09 -0700 Subject: [PATCH 06/11] skill: quest skills report through the shared decisions skill Co-Authored-By: Claude Opus 5.5 --- .claude/shared | 2 +- .claude/skills/decisions | 1 + .claude/skills/spawn-quests/SKILL.md | 7 ++----- .claude/skills/start-quest/SKILL.md | 3 +-- 4 files changed, 5 insertions(+), 8 deletions(-) create mode 120000 .claude/skills/decisions diff --git a/.claude/shared b/.claude/shared index 763fb4c9865..c0652fc80c1 160000 --- a/.claude/shared +++ b/.claude/shared @@ -1 +1 @@ -Subproject commit 763fb4c9865a0625716de8cde80d4af1b5edcef7 +Subproject commit c0652fc80c17770b4c28f310e1ba2b04817b3a2f diff --git a/.claude/skills/decisions b/.claude/skills/decisions new file mode 120000 index 00000000000..b712a8f617f --- /dev/null +++ b/.claude/skills/decisions @@ -0,0 +1 @@ +../shared/decisions \ No newline at end of file diff --git a/.claude/skills/spawn-quests/SKILL.md b/.claude/skills/spawn-quests/SKILL.md index cfcf91e6d4b..2c326476b54 100644 --- a/.claude/skills/spawn-quests/SKILL.md +++ b/.claude/skills/spawn-quests/SKILL.md @@ -29,11 +29,8 @@ Other sessions share this machine: hold new agents while the load average exceed Each agent opens its PR as a draft, lists every open decision in its report with a recommendation, and never marks it ready itself. Report each sub-agent's final status, staying silent on interim notifications, but do not monitor their PRs. -As each agent reports, explain its result in a few lines, then prompt the user inline without waiting for the rest. -Ask about its decisions (naming, API shape, branch, blockers, manual steps), each with the PR, a short summary, and your recommendation, then offer its follow-ups as a multi-select to /plan-quests. -When nothing is open, ask whether to mark the PR ready. -Resume the owning agent with each answer, and prompt again on anything its next report raises. -Mark the PR ready yourself once all of its decisions are resolved. +As each agent reports, follow /decisions without waiting for the rest, offering its follow-ups to /plan-quests. +Resume the owning agent with each answer, and mark its PR ready yourself once all of its decisions are resolved. Run /plan-quests for any selected quests in the foreground. Perform any research and monitoring in the background. diff --git a/.claude/skills/start-quest/SKILL.md b/.claude/skills/start-quest/SKILL.md index 9b9742fc1a5..1cc876c28e1 100644 --- a/.claude/skills/start-quest/SKILL.md +++ b/.claude/skills/start-quest/SKILL.md @@ -17,6 +17,5 @@ Implement the quest until it is complete, or some blocker is hit, then create a Keep scratch files (PR body, logs, notes) in the worktree's gitignored `.scratch/`. Never write to or clean up a directory other agents share, such as a session scratchpad. -When done, explain the result in a few lines and summarize any issues encountered. -List every open decision (naming, API shape, branch, blockers, manual steps) with a recommendation, and prompt the user interactively when you can, offering follow-ups as a multi-select to /plan-quests. +When done, summarize any issues encountered and follow /decisions, offering follow-ups to /plan-quests. Keep the PR a draft until the user has confirmed every decision, then mark it ready for review. From 1e5f778e44b3923c047b545dfcd0da04c109cac8 Mon Sep 17 00:00:00 2001 From: Luke Curley Date: Sat, 26 Sep 2026 14:00:27 -0700 Subject: [PATCH 07/11] chore: pin shared skills to main with decisions Co-Authored-By: Claude Opus 5.5 --- .claude/shared | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/.claude/shared b/.claude/shared index c0652fc80c1..2ee71f2abd6 160000 --- a/.claude/shared +++ b/.claude/shared @@ -1 +1 @@ -Subproject commit c0652fc80c17770b4c28f310e1ba2b04817b3a2f +Subproject commit 2ee71f2abd66b9f64d141a0342c8ab146e40cbe9 From cc15c1b869e809935e5aa1b09d0fff718d1e4285 Mon Sep 17 00:00:00 2001 From: Luke Curley Date: Sat, 26 Sep 2026 14:08:19 -0700 Subject: [PATCH 08/11] skill: move the quest hand-off to kixelated/quest, pin shared skills with it inlined Co-Authored-By: Claude Opus 5.5 --- .claude/shared | 2 +- .claude/skills/decisions | 1 - .claude/skills/spawn-quests/SKILL.md | 5 +---- .claude/skills/start-quest/SKILL.md | 5 +++-- 4 files changed, 5 insertions(+), 8 deletions(-) delete mode 120000 .claude/skills/decisions diff --git a/.claude/shared b/.claude/shared index 2ee71f2abd6..dc66c291c53 160000 --- a/.claude/shared +++ b/.claude/shared @@ -1 +1 @@ -Subproject commit 2ee71f2abd66b9f64d141a0342c8ab146e40cbe9 +Subproject commit dc66c291c539787a98b1ccb3779f1a2d71491cf4 diff --git a/.claude/skills/decisions b/.claude/skills/decisions deleted file mode 120000 index b712a8f617f..00000000000 --- a/.claude/skills/decisions +++ /dev/null @@ -1 +0,0 @@ -../shared/decisions \ No newline at end of file diff --git a/.claude/skills/spawn-quests/SKILL.md b/.claude/skills/spawn-quests/SKILL.md index 2c326476b54..b8e0f44ffab 100644 --- a/.claude/skills/spawn-quests/SKILL.md +++ b/.claude/skills/spawn-quests/SKILL.md @@ -26,11 +26,8 @@ Each agent blocks on its own checks and reports back only when done or blocked. Limit the concurrency to at most N agents in parallel, where N is half the number of physical CPU cores. Other sessions share this machine: hold new agents while the load average exceeds the core count. -Each agent opens its PR as a draft, lists every open decision in its report with a recommendation, and never marks it ready itself. Report each sub-agent's final status, staying silent on interim notifications, but do not monitor their PRs. - -As each agent reports, follow /decisions without waiting for the rest, offering its follow-ups to /plan-quests. -Resume the owning agent with each answer, and mark its PR ready yourself once all of its decisions are resolved. +Prompt the user if they want to /plan-quests for any suggested follow-ups. Run /plan-quests for any selected quests in the foreground. Perform any research and monitoring in the background. diff --git a/.claude/skills/start-quest/SKILL.md b/.claude/skills/start-quest/SKILL.md index 1cc876c28e1..73d0911e40b 100644 --- a/.claude/skills/start-quest/SKILL.md +++ b/.claude/skills/start-quest/SKILL.md @@ -17,5 +17,6 @@ Implement the quest until it is complete, or some blocker is hit, then create a Keep scratch files (PR body, logs, notes) in the worktree's gitignored `.scratch/`. Never write to or clean up a directory other agents share, such as a session scratchpad. -When done, summarize any issues encountered and follow /decisions, offering follow-ups to /plan-quests. -Keep the PR a draft until the user has confirmed every decision, then mark it ready for review. +When done, summarize any issues encounted, and suggest potential follow-up. +If you're happy with the outcome, switch the draft PR to ready for review. +If you want another set of eyes on it, keep it a draft. From 0ba6265da2f9105c626ad73abfcfe1a59963e35d Mon Sep 17 00:00:00 2001 From: Luke Curley Date: Sat, 26 Sep 2026 14:23:58 -0700 Subject: [PATCH 09/11] docs(contributing): document where each package's version lives, including OBS The generic bump skill reads it from here now. Co-Authored-By: Claude Opus 5.5 --- .claude/shared | 2 +- CONTRIBUTING.md | 11 +++++++++++ 2 files changed, 12 insertions(+), 1 deletion(-) diff --git a/.claude/shared b/.claude/shared index dc66c291c53..0e860b549d9 160000 --- a/.claude/shared +++ b/.claude/shared @@ -1 +1 @@ -Subproject commit dc66c291c539787a98b1ccb3779f1a2d71491cf4 +Subproject commit 0e860b549d937ba6318c173ccdc7fc0da2991962 diff --git a/CONTRIBUTING.md b/CONTRIBUTING.md index 10c8135fdb7..3b0c3fa16df 100644 --- a/CONTRIBUTING.md +++ b/CONTRIBUTING.md @@ -57,3 +57,14 @@ For non-trivial tasks, file an issue or offer to run `/plan-quests`. Its `moq-sync` workflow merges n0-computer/noq weekly as a PR; review it like any other, and `PARENT` names the upstream commit each release includes. A carried change lists its upstream PR, or the reason it has none, in the fork PR. For an advisory against noq or Quinn, compare the pinned release's `PARENT` with the fixing upstream commit, then sync, release the fork, and bump the pin here. + +# Versions + +Releases are cut separately; bump only when asked. Each package's version lives in one place: + +- **Rust**: release-plz owns crate versions and Rust dependency requirements. +- **JavaScript**: `js/*/package.json` packages with a `scripts.release` entry (skip private ones like `@moq/clock`, `@moq/wasm`), plus the matching workspace version in `bun.lock`. +- **Python**: `py/moq-rs/pyproject.toml` only; `py/moq-ffi` follows the `moq-ffi-v*` tag and Rust crate. +- **Swift**: `swift/VERSION`. **Kotlin**: `moq.version` in `kt/gradle.properties`. Their FFI counterparts track the Rust crate. +- **Go**: `go/wrapper/VERSION` holds a human-owned `MAJOR.MINOR` line; CI derives the patch, so only edit it for a breaking API. Leave the placeholder FFI version in `go.mod` alone. +- **OBS**: the plugin takes the `libmoq` crate's version (`cpp/obs/CMakeLists.txt`), so release-plz owns it. From f138ff16676dd88cbab3ce66834db5a511070230 Mon Sep 17 00:00:00 2001 From: Luke Curley Date: Sat, 26 Sep 2026 15:02:46 -0700 Subject: [PATCH 10/11] chore: pin shared skills with the PR paper-trail rule Co-Authored-By: Claude Opus 5.5 --- .claude/shared | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/.claude/shared b/.claude/shared index 0e860b549d9..e12b7a9dee4 160000 --- a/.claude/shared +++ b/.claude/shared @@ -1 +1 @@ -Subproject commit 0e860b549d937ba6318c173ccdc7fc0da2991962 +Subproject commit e12b7a9dee41b1faf8b3d4344e4233d3f5311818 From 9fbbf38c49e0ca6f3e1c323473302dffaff89a40 Mon Sep 17 00:00:00 2001 From: Luke Curley Date: Sat, 26 Sep 2026 18:02:14 -0700 Subject: [PATCH 11/11] docs: list Dart version source; Go VERSION bumps only on a breaking change Co-Authored-By: Claude Opus 5.5 --- CONTRIBUTING.md | 2 +- go/wrapper/README.md | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/CONTRIBUTING.md b/CONTRIBUTING.md index 3b0c3fa16df..eed93b15321 100644 --- a/CONTRIBUTING.md +++ b/CONTRIBUTING.md @@ -65,6 +65,6 @@ Releases are cut separately; bump only when asked. Each package's version lives - **Rust**: release-plz owns crate versions and Rust dependency requirements. - **JavaScript**: `js/*/package.json` packages with a `scripts.release` entry (skip private ones like `@moq/clock`, `@moq/wasm`), plus the matching workspace version in `bun.lock`. - **Python**: `py/moq-rs/pyproject.toml` only; `py/moq-ffi` follows the `moq-ffi-v*` tag and Rust crate. -- **Swift**: `swift/VERSION`. **Kotlin**: `moq.version` in `kt/gradle.properties`. Their FFI counterparts track the Rust crate. +- **Swift**: `swift/VERSION`. **Kotlin**: `moq.version` in `kt/gradle.properties`. **Dart**: `version` in `dart/moq/pubspec.yaml`. Their FFI counterparts track the Rust crate. - **Go**: `go/wrapper/VERSION` holds a human-owned `MAJOR.MINOR` line; CI derives the patch, so only edit it for a breaking API. Leave the placeholder FFI version in `go.mod` alone. - **OBS**: the plugin takes the `libmoq` crate's version (`cpp/obs/CMakeLists.txt`), so release-plz owns it. diff --git a/go/wrapper/README.md b/go/wrapper/README.md index cf241671b54..dbc9bd15b07 100644 --- a/go/wrapper/README.md +++ b/go/wrapper/README.md @@ -115,7 +115,7 @@ deliver them, and there is no stream fallback. ## Versioning -`VERSION` holds the human-owned `MAJOR.MINOR` line (the wrapper API version). Bump it in a PR when the wrapper's own API changes. The patch number is derived by CI from the existing mirror tags, so every release (whether triggered by a wrapper change or by a new `moq.dev/moq-ffi`) just takes the next patch on that line. +`VERSION` holds the human-owned `MAJOR.MINOR` line (the wrapper API version). Bump it in a PR only for a breaking change to the wrapper's own API. The patch number is derived by CI from the existing mirror tags, so every release (whether triggered by a wrapper change or by a new `moq.dev/moq-ffi`) just takes the next patch on that line. The committed `go.mod` carries a `require moq.dev/moq-ffi v0.0.0` **placeholder**. Do not "fix" it or add a `replace`: `just go check` injects a local `replace` to the freshly-generated bindings, and CI rewrites the `require` to the latest published `moq.dev/moq-ffi` at release time. Because Go resolves to the maximum version across the build graph, that `require` is a floor. Consumers always get an ffi at least as new as the wrapper was built against.