@tekyzinc/gsd-t 2.10.3 → 2.11.1

package/CHANGELOG.md

@@ -2,6 +2,20 @@
 
 All notable changes to GSD-T are documented here. Updated with each release.
 
+## [2.11.1] - 2026-02-12
+
+### Changed
+- `gsd-t-resume` now detects same-session vs cross-session mode — skips full state reload when context is already available, auto-resumes at Level 3
+- Added "Conversation vs. Work" rule to global CLAUDE.md template — plain text questions are answered conversationally, workflow only runs when a `/gsd-t-*` command is invoked
+
+## [2.11.0] - 2026-02-12
+
+### Added
+- Autonomy-level-aware auto-advancing for all phase commands — at Level 3 (Full Auto), partition, plan, impact, execute, test-sync, integrate, verify, and complete-milestone auto-advance without waiting for user input
+- Wave error recovery auto-remediates at Level 3 (up to 2 fix attempts before stopping)
+- Discuss phase always pauses for user input regardless of autonomy level
+- Autonomy levels documentation added to GSD-T-README Configuration section
+
 ## [2.10.3] - 2026-02-11
 
 ### Changed

package/commands/gsd-t-execute.md

@@ -123,6 +123,11 @@ When all tasks in all domains are complete:
 1. Update `.gsd-t/progress.md` — all tasks marked complete
 2. Set status to `EXECUTED`
 3. List any contract deviations or decisions made during execution
-4. Recommend: proceed to integrate phase
+
+### Autonomy Behavior
+
+**Level 3 (Full Auto)**: Log a brief status line (e.g., "✅ Execute complete — {N}/{N} tasks done") and auto-advance to the next phase. Do NOT wait for user input.
+
+**Level 1–2**: Report completion summary and recommend proceeding to integrate phase. Wait for confirmation.
 
 $ARGUMENTS

package/commands/gsd-t-impact.md

@@ -211,9 +211,10 @@ Validate the test landscape before recommending proceed/block:
 ### If PROCEED WITH CAUTION:
 "⚠️ Impact analysis found {N} items requiring attention:"
 - List the yellow items
-- "These can be addressed during execution. Proceed?"
-- If user confirms, continue
-- If user declines, pause for remediation
+
+**Level 3 (Full Auto)**: Log the caution items and auto-advance to execute. Do NOT wait for user input.
+
+**Level 1–2**: "These can be addressed during execution. Proceed?" Wait for user confirmation. If user declines, pause for remediation.
 
 ### If BLOCK:
 "🛑 Impact analysis found breaking changes that must be addressed first:"

package/commands/gsd-t-integrate.md

@@ -129,8 +129,13 @@ Update `.gsd-t/progress.md`:
 - Add contract audit results
 - Add smoke test results
 - List any new tasks created for gaps
-- Recommend: proceed to verify phase
 
 Commit: `[integration] Wire domains together — all contracts verified`
 
+### Autonomy Behavior
+
+**Level 3 (Full Auto)**: Log a brief status line (e.g., "✅ Integrate complete — all domain boundaries wired, {N} contracts verified") and auto-advance to the next phase. Do NOT wait for user input.
+
+**Level 1–2**: Report integration results and recommend proceeding to verify phase. Wait for confirmation.
+
 $ARGUMENTS

package/commands/gsd-t-partition.md

@@ -184,6 +184,10 @@ Before finishing, verify:
 - [ ] Every contract has an owner and at least one consumer
 - [ ] Integration checkpoints are identified for all cross-domain dependencies
 
-Report the partition to the user with a summary of domains, contracts, and any decisions that need input.
+### Autonomy Behavior
+
+**Level 3 (Full Auto)**: Log a brief status line (e.g., "✅ Partition complete — {N} domains defined, {N} contracts written") and auto-advance to the next phase. Do NOT wait for user input.
+
+**Level 1–2**: Report the partition to the user with a summary of domains, contracts, and any decisions that need input. Wait for confirmation before proceeding.
 
 $ARGUMENTS

package/commands/gsd-t-plan.md

@@ -127,7 +127,11 @@ Update `.gsd-t/progress.md`:
 
 ## Step 6: Report
 
-Present to the user:
+### Autonomy Behavior
+
+**Level 3 (Full Auto)**: Log a brief status line (e.g., "✅ Plan complete — {N} tasks across {N} domains, {execution mode}") and auto-advance to the next phase. Do NOT wait for user input.
+
+**Level 1–2**: Present to the user:
 1. Task count per domain
 2. Dependency graph (which domains block which)
 3. Recommended execution mode:
@@ -136,4 +140,6 @@ Present to the user:
    - **Team parallel**: 15+ tasks with 3+ independent starting points
 4. Any ambiguities found during planning that need user input
 
+Wait for confirmation before proceeding.
+
 $ARGUMENTS

package/commands/gsd-t-resume.md

@@ -1,8 +1,17 @@
 # GSD-T: Resume — Continue From Last State
 
-You are restoring context and continuing from where the last session left off. This is the first command to run when starting a new Claude Code session on an existing GSD-T project.
+You are resuming work after an interruption. This handles both same-session pauses (user pressed Esc to interject) and cross-session recovery (new Claude Code session).
 
-## Step 1: Load Full State
+## Step 0: Detect Resume Mode
+
+**Same-session** (conversation context still available — you can see prior messages about the active phase/task):
+- Skip to Step 2 — you already have the context loaded
+- Do NOT re-read all state files
+
+**Cross-session** (first command in a new session, no prior conversation context):
+- Run Step 1 to load full state
+
+## Step 1: Load Full State (cross-session only)
 
 Read in this exact order:
 1. `CLAUDE.md` — project context and conventions
@@ -16,16 +25,23 @@ Read in this exact order:
 
 ## Step 2: Determine Current Position
 
-From progress.md, identify:
+From progress.md (or conversation context if same-session), identify:
 - Current milestone and status
 - Which phase we're in
 - Which tasks are done, in progress, or blocked
 - Any pending decisions or user-input-needed items
 - Last entry in the Decision Log
 
-## Step 3: Report Context
+## Step 3: Report and Continue
+
+**Level 3 (Full Auto)**: Log a brief status line and auto-resume from the current task/phase. Do NOT wait for user input.
+
+```
+🔄 Resuming: {milestone name} — {phase} — {next task or action}
+```
+
+**Level 1–2**: Present fuller context and wait for confirmation:
 
-Present to the user:
 ```
 🔄 GSD-T Resuming: {milestone name}
 Phase: {current phase}
@@ -44,7 +60,7 @@ Ready to continue? Or run /user:gsd-t-status for full details.
 ## Step 4: Continue
 
 If $ARGUMENTS specifies what to do next, proceed with that.
-Otherwise, recommend the logical next action based on current state:
+Otherwise, pick up from the logical next action based on current state:
 - Mid-execution → Continue with next unblocked task
 - Between phases → Start next phase
 - Blocked → Explain what's needed to unblock

package/commands/gsd-t-test-sync.md

@@ -296,4 +296,10 @@ Action Required:
 Generated 5 test tasks → added to current domain
 ```
 
+### Autonomy Behavior
+
+**Level 3 (Full Auto)**: Log the summary and auto-advance to the next phase. If there are failing tests, attempt auto-fix (up to 2 attempts) before continuing. Do NOT wait for user input.
+
+**Level 1–2**: Present the full report and wait for user input before proceeding.
+
 $ARGUMENTS

package/commands/gsd-t-verify.md

@@ -137,7 +137,14 @@ Update `.gsd-t/progress.md`:
 - If FAIL: Set status to `VERIFY-FAILED`, list required remediations
 - Record verification date and summary
 
-Recommend next step:
+### Autonomy Behavior
+
+**Level 3 (Full Auto)**:
+- VERIFIED → Log "✅ Verify complete — all quality gates passed" and auto-advance to complete-milestone. Do NOT wait for user input.
+- CONDITIONAL PASS → Log warnings, treat as VERIFIED, and auto-advance. Do NOT wait for user input.
+- FAIL → Auto-execute remediation tasks (up to 2 fix attempts). If still failing after 2 attempts, STOP and report to user.
+
+**Level 1–2**:
 - VERIFIED → Milestone complete, proceed to next milestone or ship
 - CONDITIONAL PASS → User decides if warnings are acceptable
 - FAIL → Return to execute phase for remediation tasks

package/commands/gsd-t-wave.md

@@ -91,8 +91,16 @@ Work through each phase that hasn't been completed:
 Between each phase:
 1. Update `.gsd-t/progress.md`
 2. Report brief status to user
-3. If any phase produces findings that need user input, STOP and ask
-4. If all clear, continue to next phase
+
+### Autonomy Behavior
+
+**Level 3 (Full Auto)**: Auto-advance to the next phase after logging status. Only STOP for:
+- Destructive Action Guard violations (always)
+- Impact analysis BLOCK verdict (always)
+- Unrecoverable errors after 2 fix attempts
+- The Discuss phase (always pauses for user input, even at Level 3)
+
+**Level 1–2**: If any phase produces findings that need user input, STOP and ask. If all clear, continue to next phase.
 
 Status messages:
 ```
@@ -147,25 +155,28 @@ If the user interrupts or the session needs to end:
 - Report blocking issues
 - Generate remediation tasks
 - Add to appropriate domain
-- Ask: "Address blockers now, or pause?"
-- If address: execute remediation tasks, re-run impact
-- If pause: save state, exit
+
+**Level 3 (Full Auto)**: Auto-execute remediation tasks, then re-run impact analysis. Only STOP if remediation fails after 2 attempts.
+
+**Level 1–2**: Ask: "Address blockers now, or pause?" If address: execute remediation tasks, re-run impact. If pause: save state, exit.
 
 ### If tests fail during execute:
 - Pause execution
 - Report failing tests
 - Generate fix tasks
-- Ask: "Fix now or continue?"
-- If fix: execute fix tasks, re-run tests
-- If continue: note failures, proceed (will catch in verify)
+
+**Level 3 (Full Auto)**: Auto-execute fix tasks and re-run tests (up to 2 fix attempts). If still failing, STOP and report to user.
+
+**Level 1–2**: Ask: "Fix now or continue?" If fix: execute fix tasks, re-run tests. If continue: note failures, proceed (will catch in verify).
 
 ### If verify fails:
 - Report failures
 - Generate remediation tasks
 - Do NOT run complete-milestone
-- Ask: "Address issues now?"
-- If yes: execute remediation, re-run verify
-- If no: save state with VERIFY_FAILED status
+
+**Level 3 (Full Auto)**: Auto-execute remediation tasks and re-run verify (up to 2 attempts). If still failing, STOP and report to user.
+
+**Level 1–2**: Ask: "Address issues now?" If yes: execute remediation, re-run verify. If no: save state with VERIFY_FAILED status.
 
 ## Workflow Visualization
 

package/docs/GSD-T-README.md

@@ -250,6 +250,16 @@ Verify with: `/user:gsd-t-help`
 
 GSD-T respects your project's `CLAUDE.md` for conventions and autonomy level.
 
+### Autonomy Levels
+
+| Level | Phase Behavior |
+|-------|---------------|
+| **Level 1: Supervised** | Pauses at each phase for confirmation |
+| **Level 2: Standard** | Pauses at milestones |
+| **Level 3: Full Auto** (default) | Auto-advances through all phases. Only stops for: Destructive Action Guard, impact BLOCK verdicts, unrecoverable errors (after 2 fix attempts), and the Discuss phase |
+
+Set in your project's `CLAUDE.md` under `## Autonomy Level`.
+
 Recommended `.gsd-t/config.json`:
 
 ```json

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@tekyzinc/gsd-t",
-  "version": "2.10.3",
+  "version": "2.11.1",
   "description": "GSD-T: Contract-Driven Development for Claude Code — 35 slash commands with backlog management, impact analysis, test sync, and milestone archival",
   "author": "Tekyz, Inc.",
   "license": "MIT",

package/templates/CLAUDE-global.md

@@ -149,6 +149,10 @@ Even in development, the user may have:
 
 # Autonomous Execution Rules
 
+## Conversation vs. Work
+
+Only execute GSD-T workflow behavior when a `/gsd-t-*` command is invoked or when actively mid-phase (resumed via `/gsd-t-resume`). **Plain text messages — especially questions — should be answered conversationally.** Do not launch into workflow execution, file reading, or phase advancement from a question or comment. If the user wants work done, they will invoke a command.
+
 ## Prime Rule
 KEEP GOING. Only stop for:
 1. Unrecoverable errors after 2 fix attempts