* fix(auto): add missing import for resolveSkillDiscoveryMode
Used at line 687 but not imported, causing "resolveSkillDiscoveryMode is
not defined" crash on auto-mode startup.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
* fix(auto): add workingDirectory to all auto-mode prompt templates
Six prompt templates (reassess-roadmap, complete-milestone, replan-slice,
run-uat, research-milestone, plan-milestone) were missing the working
directory directive. Without it, the LLM infers the main repo path from
system context and cd's there instead of staying in the worktree. This
causes artifacts to be written to the wrong location, preventing the
dispatch loop from detecting completion and triggering infinite
re-dispatches of the same unit.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
* fix(auto): detect mid-session resource updates and stop gracefully
Templates are read from disk on each dispatch but extension code is
loaded once at startup. If resources are re-synced mid-session (via
/gsd:update, npm update, or dev copy-resources), templates may expect
variables the in-memory code doesn't provide, causing a crash.
Add a syncedAt timestamp to managed-resources.json. Auto-mode captures
this at startup and checks before each dispatch. If resources changed,
it stops with a clear message instead of crashing.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
* test: add workingDirectory to prompt template test fixtures
Tests that load prompt templates via loadPromptFromWorktree now pass the
workingDirectory variable, matching the updated templates that include
the {{workingDirectory}} directive.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
After deleting summary files and modifying PLAN files, only
invalidateStateCache() was called. Path and parse caches remained
stale, causing deriveState() to return incorrect results — showing
undone tasks as still complete.
Four fixes to auto-recovery logic that caused silent failures or
inconsistent state:
1. skipExecuteTask: return false when checkbox regex doesn't match the
plan format, so callers fall through to other recovery strategies
instead of assuming success (lines 252-255)
2. verifyExpectedArtifact: fail verification on corrupt/unparseable
roadmap instead of silently passing. Prevents advancing past an
incomplete complete-slice when the roadmap file is malformed (line 152)
3. removePersistedKey: use atomic tmp+rename write (matching
persistCompletedKey) to prevent completed-units.json corruption
on crash mid-write (line 293)
4. selfHealRuntimeRecords: use verifyExpectedArtifact instead of bare
existsSync for execute-task healing, so tasks with summary but
unchecked plan checkbox aren't incorrectly marked complete (line 374)
Co-authored-by: TÂCHES <afromanguy@me.com>
The progress bar in the auto-mode widget was snapshot-based — only
updated at dispatch time via updateSliceProgressCache(). During
long-running units (especially after the worktree architecture in
PR #506), the bar appeared frozen even as tasks completed on disk.
Add a 5-second interval inside the widget that re-reads the roadmap
and plan files from disk, so slice/task progress reflects reality
without waiting for the next unit dispatch.
Closes#549
#536 changed git.isolation from deprecated to an active setting.
Update the test to verify it passes through correctly instead of
expecting a deprecation warning. Add separate test for the still-
deprecated git.merge_to_main.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Introduce typed error hierarchy (GSDError with stable error codes) for
programmatic error matching and crash diagnostics. Convert
MergeConflictError to extend GSDError. Capture error references in the
most impactful silent catch blocks across crash-recovery, auto-recovery,
and activity-log — errors remain non-fatal but are no longer discarded.
Closes#525
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Three independent caches (state, path, parse) required manual coordination
on every dispatch cycle. Forgetting any one caused stale reads (#431).
Add a single invalidateAllCaches() in cache.ts that clears all three,
and replace grouped call sites in auto.ts and tests.
Individual clear functions are preserved for callers that legitimately
only need to clear one cache.
Closes#527
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Replace the 130-line if-else chain in dispatchNextUnit with a
declarative DispatchRule[] table in auto-dispatch.ts.
Each rule maps a GSD state to the unit type, unit ID, and prompt
builder. Rules are evaluated in order; first match wins. The table
is inspectable, testable per-rule, and extensible without modifying
orchestration code.
- auto-dispatch.ts: 258 lines, 12 named rules
- auto.ts dispatch section: 130 lines → 20 lines
- Updated auto-draft-pause test to verify rules in new location
- 123/123 tests pass, zero TypeScript errors
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Validates parsed preferences against known keys and expected types.
Unknown keys produce warnings instead of being silently ignored.
Previously unvalidated fields (budget_enforcement, context_pause_threshold,
models, auto_supervisor, notifications, remote_questions) are now
type-checked. Warnings surface through LoadedGSDPreferences so callers
can inspect validation results.
Closes#522
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Three fixes for slice transition crashes and git isolation regression:
1. dispatch-guard reads from disk instead of git branch — prevents
false blockers when roadmap state is committed on milestone branch
but not yet on the integration branch (#530).
2. Auto-resolve .gsd/ state file conflicts during milestone merge and
in the mid-merge safety check. STATE.md, completed-units.json, and
auto.lock diverge between branches during normal operation — always
prefer the milestone branch version. Only escalate non-.gsd conflicts
to MergeConflictError (#530).
3. Restore git.isolation preference with two values (#531):
- "worktree" (default): creates milestone worktrees for isolated work
- "branch": works directly in the project root, skipping worktree
creation — for submodule-heavy repos where worktrees fail
The branchless worktree architecture remains the default. Branch mode
simply gates worktree entry points so no worktree is ever created.
Comprehensive prompt and template overhaul addressing multiple issues
discovered during auto-mode execution:
**Worktree cwd fix** — Executor agents wrote code to the main repo
instead of the worktree because prompts never stated the working
directory. Added ## Working Directory section with explicit path to
execute-task, plan-slice, research-slice, complete-slice prompts.
Passed workingDirectory: base to all loadPrompt() calls in
auto-prompts.ts and guided-flow.ts.
**Stale branch references** — Removed all "slice branch" references
(branchless since v2.14.0). Updated system.md with Worktree Model
section. Updated preferences-reference.md descriptions.
**System prompt updates** — Added REQUIREMENTS.md, CONTEXT.md docs,
system-managed directories (runtime/, activity/, worktrees/) to
directory structure.
**Pipeline awareness** — Every phase now knows its role: researchers
are scouts writing for planners, planners trust research and don't
re-explore, executors build from task plans, completers write for
downstream readers. Eliminates redundant work between phases.
**Research depth calibration** — Three-tier system (deep/targeted/light)
across research-slice, guided-research-slice, research-milestone.
Light research for known patterns can be 15-20 lines.
**Template improvements:**
- research.md: "Existing Code and Patterns" → "Implementation Landscape"
with Key Files, Build Order, Verification Approach subsections
- plan.md: reduced task examples from 3 to 2 to avoid anchoring
- state.md: removed dead Active Workspace field
- reassessment.md: added depth guidance for no-change vs modified
- Carry-forward now extracts key_files (was missing — executors
couldn't see which files prior tasks created)
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
The dispatch gap watchdog is a one-shot timer that fires 5s after a unit
completes without a follow-up dispatch. Previously, if the watchdog's
dispatchNextUnit() call returned without actually dispatching a unit
(no sendMessage called), auto-mode was left permanently active but idle
— no new watchdog was started and no stopAuto was called.
This happened when:
- State between milestones had no dispatchable unit
- Stale completed-units.json after GSD updates caused skip loops
- dispatchNextUnit silently returned without finding work
Now the watchdog checks whether a unit was actually dispatched after its
retry attempt. If not, it stops auto-mode cleanly with a user-facing
message instead of leaving it stuck.
Closes#537
The migrate/ directory (1,862 lines across 9 files) is one-time migration
code for .planning/ → .gsd/ conversion. Replace the static top-level
import with a dynamic import() that only loads when `/gsd migrate` is
invoked, matching the existing pattern used for hooks and metrics.
Closes#523
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Replace existsSync collision loop with atomic O_CREAT|O_EXCL file
creation, hoist regex to module-level constant, and memoize
getPackageDir() to avoid repeated directory walks.
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Arrow keys produce `^[[D`/`^[[C` instead of moving the cursor when event
loop latency causes the StdinBuffer to split escape sequences.
Three layered fixes:
1. Increase StdinBuffer timeout from 10ms to 50ms (matches xterm default)
so split escape sequences are reassembled even under load.
2. Clean up stale readline listeners after @clack/prompts onboarding —
readline.emitKeypressEvents() leaves a permanent data listener that
is unnecessary for the TUI.
3. Guard in editor against CSI remnants: if a split still occurs, reject
text matching navigation escape patterns ([A-F, [H, [Z, [n~) instead
of inserting them as characters.
Closes#493
When auto-mode creates a worktree and chdir's into it, the Node process
cwd changes but AgentSession._cwd stays frozen at the original path.
Every newSession() builds a system prompt telling the LLM "Current
working directory: /original/path", so the LLM cd's back there and
writes files to the wrong location.
Update _cwd = process.cwd() at the start of newSession() so the system
prompt reflects the actual working directory after chdir.
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Auto-worktrees are fresh git checkouts — untracked .gsd/ files don't
carry over. Projects with the old blanket .gsd/ gitignore have planning
artifacts on disk but not in git. When createAutoWorktree makes a new
worktree, the milestones/, DECISIONS.md, REQUIREMENTS.md etc are missing,
causing auto-mode to loop on plan-slice (plan file not found in worktree).
Copy .gsd/ planning artifacts from the source repo into the new worktree
after git worktree add. Skips runtime files and the worktrees/ dir.
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
ensureGitignore() now detects and removes standalone ".gsd/" lines that
blanket-ignore the entire directory. Replaces with explicit runtime-only
patterns so .gsd/milestones/ planning artifacts are tracked in git.
Without this, existing projects keep the old blanket ignore forever.
New worktrees start with zero planning state because artifacts aren't
in git, causing auto-mode to re-execute completed work.
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- auto.ts: wrap dispatchNextUnit body in try/finally to always reset
_dispatching to false. Without this, the reentrancy guard permanently
blocked all subsequent dispatches after the first one, causing the
dispatch gap watchdog to fire and auto-mode to stall.
- discuss.md: render depth summary as chat text (where markdown renders)
then use ask_user_questions for the short confirmation only.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Downgrade internal recovery machinery to info/verbose-only so users
only see warnings when action is needed:
- "Dispatch gap detected" → verbose-only info (recovery is automatic)
- "Model not found, trying fallback" → verbose-only info
- "Failed to set model, trying fallback" → verbose-only info
- "Could not set any preferred model" → deleted (redundant)
- "New session cancelled" → info (user action, not error)
- "Unexpected phase" → info with doctor suggestion
- "No command context" → info with restart suggestion
Kept as warnings (user-actionable):
- Budget ceiling, blockers, prior slice incomplete, pre-flight,
no context, stub summary, model ambiguity, all fallbacks exhausted
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
