releases/ClaudeDo
2
0
Fork 0
Code Issues 1 Pull Requests 2 Actions Packages Projects Releases 24 Wiki Activity

fix(worker): transport ConPTY task brief via file, not CLI argument

Browse Source 
BuildForTaskAsync's fresh-session path flattened the task title+description
into one positional CLI argument, which the ConPTY host joins into a single
command line and claude re-splits on whitespace -- any dash-leading token in
the description (e.g. "->", "--abort") was misread as an unknown option, and
a raw multi-line prompt truncated at its first newline regardless. Now the
brief is written to ~/.todo-app/task-sessions/<taskId>/brief.md and exposed
via --add-dir, with a single-line kickoff pointing claude at it -- the same
pattern BuildForMergeHelperAsync and the planning launcher already use.
This commit is contained in:
mika kuns
2026-08-05 10:54:48 +02:00
parent 334cf1e1d2
commit 3972ce50a6
3 changed files with 129 additions and 18 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/ClaudeDo.Worker/CLAUDE.md
+1 -1
View File
@@ -157,7 +157,7 @@ Each CLI invocation is recorded in the `task_runs` table via `TaskRunRepository`
- Review/merge: `ApproveReview(taskId, targetBranch) -> MergeResultDto` (childless task: merges its worktree then Done, conflict stays WaitingForReview; task with children: drives `PlanningMergeOrchestrator` to merge the whole unit), `ContinuePlanningMerge` / `AbortPlanningMerge` (resolve a unit-merge conflict), `PreviewMerge(taskId, targetBranch) -> MergePreviewDto` (non-destructive mergeability check), `RejectReviewToQueue`, `RejectReviewToIdle`, `CancelReview`, `MergeTask`, `GetMergeTargets`
- Single-task conflict resolver (Layer C): `StartConflictMerge`, `GetMergeConflictDocuments` (segments), `WriteConflictResolution`, `ContinueConflictMerge`, `AbortConflictMerge` (service-level `TaskMergeService.ContinueMergeAsync`/`AbortMergeAsync` keep their names)
- Planning sessions: `StartPlanningSession`, `ResumePlanningSession`, `DiscardPlanningSession`, `FinalizePlanningSession`, `QueuePlanningSubtasks`, `GetPendingDraftCount`, `GetPlanningAggregate` (per-subtask diffs), `BuildPlanningIntegrationBranch` (combined diff)
- Interactive sessions (embedded ConPTY, UI process): `GetInteractiveLaunchSpec`, `GetAdHocLaunchSpec`, `GetMergeHelperLaunchSpec`, `CreateMergeHelperTask` (creates the ClaudeDo task that owns a list-handler run — `Idle`/`IsManual=true`, `HandlerBaseCommit` stamped to the list repo's current HEAD via `InteractiveLaunchSpecService.CreateMergeHelperTaskAsync` — called by the UI before it opens the task-based ConPTY tile), `SubmitTaskForReview` (branches on whether the task has a `WorktreeEntity`: with one, commits it and moves on; without one, it's a worktree-less list-handler host task and it just stamps `HandlerHeadCommit` to the list repo's current HEAD — both paths then flip the task Idle/Failed → WaitingForReview). **Every** ConPTY spec that `InteractiveLaunchSpecService` builds leads with `--effort <level>` from the relevant model's preset (task/list model for a task session, `PlanningAlias` for planning, list config for the list handler, global default for ad-hoc). `--model` is deliberately NOT forced on an interactive session — the user can still switch models in the TUI. The list-handler spec (`BuildForMergeHelperAsync`) uses `--permission-mode auto` so it runs unattended; the `--allowedTools` allowlist (`mcp__claudedo__*,Read,Grep,Glob,Edit,Bash,WebFetch,WebSearch,Skill`) remains the security boundary.
- Interactive sessions (embedded ConPTY, UI process): `GetInteractiveLaunchSpec`, `GetAdHocLaunchSpec`, `GetMergeHelperLaunchSpec`, `CreateMergeHelperTask` (creates the ClaudeDo task that owns a list-handler run — `Idle`/`IsManual=true`, `HandlerBaseCommit` stamped to the list repo's current HEAD via `InteractiveLaunchSpecService.CreateMergeHelperTaskAsync` — called by the UI before it opens the task-based ConPTY tile), `SubmitTaskForReview` (branches on whether the task has a `WorktreeEntity`: with one, commits it and moves on; without one, it's a worktree-less list-handler host task and it just stamps `HandlerHeadCommit` to the list repo's current HEAD — both paths then flip the task Idle/Failed → WaitingForReview). Every ConPTY spec that `InteractiveLaunchSpecService` builds passes `--effort <level>` from the relevant model's preset (task/list model for a task session, `PlanningAlias` for planning, list config for the list handler, global default for ad-hoc) — it leads the args except for a fresh task session with a brief, where `--add-dir <sessionDir>` must come first so `--effort` (a single-value flag) can sit directly before the positional kickoff (see below). `--model` is deliberately NOT forced on an interactive session — the user can still switch models in the TUI. The list-handler spec (`BuildForMergeHelperAsync`) uses `--permission-mode auto` so it runs unattended; the `--allowedTools` allowlist (`mcp__claudedo__*,Read,Grep,Glob,Edit,Bash,WebFetch,WebSearch,Skill`) remains the security boundary. **No ConPTY path ever passes task free-text (title/description/brief) as a CLI argument** — every one of them (task session, planning start/resume, list handler) writes it to a file first and hands claude a single-line kickoff pointing at that file, exposed via `--add-dir`. Reason: the ConPTY host flattens `Args` into one command line to spawn the process, and claude re-splits that line on whitespace, so any token starting with `-` (e.g. `->`, `--abort`) in real task text would be misread as an unknown option — and a raw multi-line positional prompt truncates at its first newline regardless. A fresh task session's brief lives at `~/.todo-app/task-sessions/<taskId>/brief.md` (`InteractiveLaunchSpecService.BuildFreshTaskArgsAsync`); a task with neither title nor description skips the file and the positional arg entirely.
- Worktrees: `CleanupFinishedWorktrees`, `ResetAllWorktrees`, `GetWorktreesOverview`, `SetWorktreeState`, `ForceRemoveWorktree`
- Agents/settings/lists: `GetAgents`, `RefreshAgents`, `RestoreDefaultAgents`, `GetAppSettings`, `UpdateAppSettings`, `UpdateList`, `UpdateListConfig`, `GetListConfig`, `UpdateTaskAgentSettings`
- Reports/notes/prep: `GetWeekReport`, `GenerateWeekReport`, `GetDailyNotes`, `AddDailyNote`, `UpdateDailyNote`, `DeleteDailyNote`, `RunDailyPrepNow`, `ClearMyDay`, `GetLastPrepLog`, `ListPrimeSchedules`, `UpsertPrimeSchedule`, `DeletePrimeSchedule`
src/ClaudeDo.Worker/Runner/InteractiveLaunchSpecService.cs
+40 -13
View File
@@ -94,17 +94,17 @@ public sealed class InteractiveLaunchSpecService : IInteractiveLaunchSpecService
var resolvedClaude = WindowsTerminalLauncher.Resolve(_claudePath)
?? throw new InvalidOperationException($"claude executable not found: {_claudePath}");
// Resume an existing session as-is; for a fresh session, seed the interactive TUI with
// the task's prompt (title + description) as claude's positional prompt so it starts on
// the task -- the user then supervises/answers rather than retyping it.
var args = run?.SessionId is { Length: > 0 } sessionId
? WindowsTerminalLauncher.BuildResumeArgs(sessionId)
: BuildFreshPromptArgs(task);
// Start the hand-driven session at the effort configured for the model this task would run
// under, instead of inheriting whatever the user's global Claude Code config happens to be.
// The model itself is deliberately NOT forced here — the user can still switch it in the TUI.
args = WithEffort(args, EffortFor(globalSettings, task.Model ?? listConfig?.Model));
var effort = EffortFor(globalSettings, task.Model ?? listConfig?.Model);
// Resume an existing session as-is; for a fresh session, seed the interactive TUI with
// the task's brief (title + description) via a file, never as a positional CLI argument --
// see BuildFreshTaskArgsAsync for why.
var args = run?.SessionId is { Length: > 0 } sessionId
? WithEffort(WindowsTerminalLauncher.BuildResumeArgs(sessionId), effort)
: await BuildFreshTaskArgsAsync(task, effort, ct);
// Same run environment variable ClaudeProcess sets for every headless run: the
// AskUser MCP tool call caps at 60s unless raised, and lifting it is harmless for
@@ -324,20 +324,47 @@ public sealed class InteractiveLaunchSpecService : IInteractiveLaunchSpecService
return new AppSettingsRepository(ctx).GetAsync().GetAwaiter().GetResult();
}
// The positional prompt claude opens the interactive session on. Empty (no positional arg)
// if the task has neither a title nor a description.
private static IReadOnlyList<string> BuildFreshPromptArgs(TaskEntity task)
// The task's brief (title + description, unmodified/multi-line) must never travel as a CLI
// argument: the ConPTY host flattens Args into one command line to spawn the process, and
// claude re-splits that line on whitespace, so any token in the description that starts with
// '-' (e.g. "->", "--abort") gets misread as an unknown option, and a newline would truncate
// the brief at its first line even if it didn't. So it's written to a file instead --
// same pattern as BuildForMergeHelperAsync / WindowsTerminalLauncher.BuildPlanningStartArgs --
// and claude is pointed at it with a single-line kickoff. --add-dir exposes the session dir to
// Read; --effort (single-value) must sit directly before the positional kickoff so the
// preceding variadic --add-dir doesn't swallow the kickoff as another directory.
// No brief (task has neither a title nor a description) -> no positional arg at all.
private static async Task<IReadOnlyList<string>> BuildFreshTaskArgsAsync(TaskEntity task, string effort, CancellationToken ct)
{
var brief = BuildTaskBrief(task);
if (string.IsNullOrEmpty(brief))
return new[] { "--effort", effort };
var sessionDir = Path.Combine(Paths.AppDataRoot(), "task-sessions", task.Id);
Directory.CreateDirectory(sessionDir);
var briefPath = Path.Combine(sessionDir, "brief.md");
await File.WriteAllTextAsync(briefPath, brief, ct);
return new[]
{
"--add-dir", sessionDir,
"--effort", effort,
$"Read the file {briefPath} first. It contains the task you must work on. " +
"After reading it, begin the session as your instructions describe.",
};
}
private static string BuildTaskBrief(TaskEntity task)
{
var title = task.Title?.Trim();
var description = task.Description?.Trim();
var prompt = (string.IsNullOrEmpty(title), string.IsNullOrEmpty(description)) switch
return (string.IsNullOrEmpty(title), string.IsNullOrEmpty(description)) switch
{
(false, false) => $"{title}\n\n{description}",
(false, true) => title!,
(true, false) => description!,
_ => string.Empty,
};
return string.IsNullOrEmpty(prompt) ? Array.Empty<string>() : new[] { prompt };
}
private async Task<IReadOnlyList<string>> FilterToInstalledSkillsAsync(IReadOnlyList<string> requested, CancellationToken ct)