prizmkit 1.1.119 → 1.1.121

package/bundled/dev-pipeline/scripts/update-bug-status.py

@@ -276,6 +276,33 @@ def _set_continuation_common(status_data, args, session_id):
         status_data["last_session_id"] = session_id
 
 
+def _set_ai_error_continuation_metadata(status_data, args, session_id, reason):
+    """Preserve durable task/session state for any retryable launched AI error."""
+    if args.active_dev_branch:
+        status_data["active_dev_branch"] = args.active_dev_branch
+    if args.base_branch:
+        status_data["base_branch"] = args.base_branch
+    status_data["continuation_pending"] = True
+    status_data["continuation_reason"] = reason
+    status_data["continuation_count"] = status_data.get("continuation_count", 0) + 1
+    status_data["previous_session_id"] = session_id
+    status_data["last_fatal_error_code"] = args.last_fatal_error_code or reason
+    if args.continuation_summary_path:
+        status_data["continuation_summary_path"] = args.continuation_summary_path
+    if args.no_progress_count is not None:
+        status_data["no_progress_count"] = max(0, args.no_progress_count)
+    fingerprint = _load_progress_fingerprint(args.progress_fingerprint)
+    if fingerprint is not None:
+        status_data["last_progress_fingerprint"] = fingerprint
+    status_data["resume_from_phase"] = None
+    if session_id:
+        sessions = status_data.get("sessions", [])
+        if session_id not in sessions:
+            sessions.append(session_id)
+        status_data["sessions"] = sessions
+        status_data["last_session_id"] = session_id
+
+
 def set_context_overflow_continuation_metadata(status_data, args, session_id):
     """Record continuation metadata for a context-overflow outcome in status.json."""
     status_data["context_overflow_count"] = status_data.get("context_overflow_count", 0) + 1
@@ -456,6 +483,18 @@ def action_update(args, bug_list_path, state_dir):
 
     # Track what status we write to bug-fix-list.json
     new_status = current_list_status
+    if args.active_dev_branch:
+        bs["active_dev_branch"] = args.active_dev_branch
+    if args.base_branch:
+        bs["base_branch"] = args.base_branch
+    if args.last_fatal_error_code:
+        bs["last_fatal_error_code"] = args.last_fatal_error_code
+    if args.continuation_summary_path:
+        bs["continuation_summary_path"] = args.continuation_summary_path
+    fingerprint = _load_progress_fingerprint(args.progress_fingerprint)
+    if fingerprint is not None:
+        bs["last_progress_fingerprint"] = fingerprint
+
 
     if session_status == "success":
         bs["infra_error_count"] = 0
@@ -499,10 +538,14 @@ def action_update(args, bug_list_path, state_dir):
 
         if infra_error_count >= max_infra_retries:
             new_status = "failed"
+            bs["continuation_pending"] = False
+            bs["continuation_reason"] = session_status
+            bs["needs_attention"] = True
             if session_id:
                 bs["last_failed_session_id"] = session_id
         else:
-            new_status = "pending"
+            new_status = "in_progress"
+            _set_ai_error_continuation_metadata(bs, args, session_id, session_status)
 
         err = update_bug_in_list(bug_list_path, bug_id, new_status)
         if err:
@@ -528,27 +571,25 @@ def action_update(args, bug_list_path, state_dir):
     else:
         bs["retry_count"] = bs.get("retry_count", 0) + 1
 
-        cleaned = cleanup_bug_artifacts(
-            state_dir=state_dir,
-            bug_id=bug_id,
-            project_root=args.project_root,
-        )
-
         if bs["retry_count"] >= max_retries:
             new_status = "failed"
+            bs["continuation_pending"] = False
+            bs["continuation_reason"] = session_status
+            bs["needs_attention"] = True
         else:
-            new_status = "pending"
+            new_status = "in_progress"
+            _set_ai_error_continuation_metadata(bs, args, session_id, session_status)
 
         bs["resume_from_phase"] = None
-        bs["sessions"] = []
-        bs["last_session_id"] = None
+        if session_id:
+            bs["last_failed_session_id"] = session_id
 
         err = update_bug_in_list(bug_list_path, bug_id, new_status)
         if err:
             error_out("Failed to update .prizmkit/plans/bug-fix-list.json: {}".format(err))
             return
 
-    if session_status == "success" and session_id:
+    if session_id:
         sessions = bs.get("sessions", [])
         if session_id not in sessions:
             sessions.append(session_id)
@@ -581,8 +622,9 @@ def action_update(args, bug_list_path, state_dir):
         summary["degraded_reason"] = session_status
         summary["restart_policy"] = "finalization_retry"
     elif session_status == "infra_error":
-        summary["restart_policy"] = "infra_retry"
+        summary["restart_policy"] = "in_process_ai_error_continuation" if new_status == "in_progress" else "needs_attention"
         summary["infra_error_count"] = bs.get("infra_error_count", 0)
+        summary.update(continuation_metadata_summary(bs))
         summary["artifacts_preserved"] = True
     elif session_status == "context_overflow":
         summary["restart_policy"] = "context_overflow_continuation"
@@ -593,8 +635,9 @@ def action_update(args, bug_list_path, state_dir):
         summary.update(continuation_metadata_summary(bs))
         summary["artifacts_preserved"] = True
     elif session_status != "success":
-        summary["restart_policy"] = "full_restart"
-        summary["cleanup_performed"] = cleaned
+        summary["restart_policy"] = "in_process_ai_error_continuation" if new_status == "in_progress" else "needs_attention"
+        summary.update(continuation_metadata_summary(bs))
+        summary["artifacts_preserved"] = True
 
     print(json.dumps(summary, indent=2, ensure_ascii=False))
 

package/bundled/dev-pipeline/scripts/update-feature-status.py

@@ -381,6 +381,33 @@ def _set_continuation_common(status_data, args, session_id):
         status_data["last_session_id"] = session_id
 
 
+def _set_ai_error_continuation_metadata(status_data, args, session_id, reason):
+    """Preserve durable task/session state for any retryable launched AI error."""
+    if args.active_dev_branch:
+        status_data["active_dev_branch"] = args.active_dev_branch
+    if args.base_branch:
+        status_data["base_branch"] = args.base_branch
+    status_data["continuation_pending"] = True
+    status_data["continuation_reason"] = reason
+    status_data["continuation_count"] = status_data.get("continuation_count", 0) + 1
+    status_data["previous_session_id"] = session_id
+    status_data["last_fatal_error_code"] = args.last_fatal_error_code or reason
+    if args.continuation_summary_path:
+        status_data["continuation_summary_path"] = args.continuation_summary_path
+    if args.no_progress_count is not None:
+        status_data["no_progress_count"] = max(0, args.no_progress_count)
+    fingerprint = _load_progress_fingerprint(args.progress_fingerprint)
+    if fingerprint is not None:
+        status_data["last_progress_fingerprint"] = fingerprint
+    status_data["resume_from_phase"] = None
+    if session_id:
+        sessions = status_data.get("sessions", [])
+        if session_id not in sessions:
+            sessions.append(session_id)
+        status_data["sessions"] = sessions
+        status_data["last_session_id"] = session_id
+
+
 def set_context_overflow_continuation_metadata(status_data, args, session_id):
     """Record continuation metadata for a context-overflow outcome in status.json."""
     status_data["context_overflow_count"] = status_data.get("context_overflow_count", 0) + 1
@@ -848,6 +875,18 @@ def action_update(args, feature_list_path, state_dir):
 
     # Track what status we write to feature-list.json
     new_status = current_list_status
+    if args.active_dev_branch:
+        fs["active_dev_branch"] = args.active_dev_branch
+    if args.base_branch:
+        fs["base_branch"] = args.base_branch
+    if args.last_fatal_error_code:
+        fs["last_fatal_error_code"] = args.last_fatal_error_code
+    if args.continuation_summary_path:
+        fs["continuation_summary_path"] = args.continuation_summary_path
+    fingerprint = _load_progress_fingerprint(args.progress_fingerprint)
+    if fingerprint is not None:
+        fs["last_progress_fingerprint"] = fingerprint
+
 
     if session_status == "success":
         fs["infra_error_count"] = 0
@@ -923,10 +962,14 @@ def action_update(args, feature_list_path, state_dir):
 
         if infra_error_count >= max_infra_retries:
             new_status = "failed"
+            fs["continuation_pending"] = False
+            fs["continuation_reason"] = session_status
+            fs["needs_attention"] = True
             if session_id:
                 fs["last_failed_session_id"] = session_id
         else:
-            new_status = "pending"
+            new_status = "in_progress"
+            _set_ai_error_continuation_metadata(fs, args, session_id, session_status)
 
         err = update_feature_in_list(feature_list_path, feature_id, new_status)
         if err:
@@ -955,8 +998,12 @@ def action_update(args, feature_list_path, state_dir):
 
         if fs["retry_count"] >= max_retries:
             new_status = "failed"
+            fs["continuation_pending"] = False
+            fs["continuation_reason"] = session_status
+            fs["needs_attention"] = True
         else:
-            new_status = "pending"
+            new_status = "in_progress"
+            _set_ai_error_continuation_metadata(fs, args, session_id, session_status)
 
         fs["resume_from_phase"] = None
         if session_id:
@@ -969,7 +1016,7 @@ def action_update(args, feature_list_path, state_dir):
             error_out("Failed to update .prizmkit/plans/feature-list.json: {}".format(err))
             return
 
-    if session_status == "success" and session_id:
+    if session_id:
         sessions = fs.get("sessions", [])
         if session_id not in sessions:
             sessions.append(session_id)
@@ -1013,8 +1060,9 @@ def action_update(args, feature_list_path, state_dir):
         summary["degraded_reason"] = session_status
         summary["restart_policy"] = "finalization_retry"
     elif session_status == "infra_error":
-        summary["restart_policy"] = "infra_retry"
+        summary["restart_policy"] = "in_process_ai_error_continuation" if new_status == "in_progress" else "needs_attention"
         summary["infra_error_count"] = fs.get("infra_error_count", 0)
+        summary.update(continuation_metadata_summary(fs))
         summary["artifacts_preserved"] = True
     elif session_status == "context_overflow":
         summary["restart_policy"] = "context_overflow_continuation"
@@ -1025,7 +1073,8 @@ def action_update(args, feature_list_path, state_dir):
         summary.update(continuation_metadata_summary(fs))
         summary["artifacts_preserved"] = True
     elif session_status != "success":
-        summary["restart_policy"] = "preserve_and_retry"
+        summary["restart_policy"] = "in_process_ai_error_continuation" if new_status == "in_progress" else "needs_attention"
+        summary.update(continuation_metadata_summary(fs))
         summary["artifacts_preserved"] = True
 
     print(json.dumps(summary, indent=2, ensure_ascii=False))

package/bundled/dev-pipeline/scripts/update-refactor-status.py

@@ -275,6 +275,33 @@ def _set_continuation_common(status_data, args, session_id):
         status_data["last_session_id"] = session_id
 
 
+def _set_ai_error_continuation_metadata(status_data, args, session_id, reason):
+    """Preserve durable task/session state for any retryable launched AI error."""
+    if args.active_dev_branch:
+        status_data["active_dev_branch"] = args.active_dev_branch
+    if args.base_branch:
+        status_data["base_branch"] = args.base_branch
+    status_data["continuation_pending"] = True
+    status_data["continuation_reason"] = reason
+    status_data["continuation_count"] = status_data.get("continuation_count", 0) + 1
+    status_data["previous_session_id"] = session_id
+    status_data["last_fatal_error_code"] = args.last_fatal_error_code or reason
+    if args.continuation_summary_path:
+        status_data["continuation_summary_path"] = args.continuation_summary_path
+    if args.no_progress_count is not None:
+        status_data["no_progress_count"] = max(0, args.no_progress_count)
+    fingerprint = _load_progress_fingerprint(args.progress_fingerprint)
+    if fingerprint is not None:
+        status_data["last_progress_fingerprint"] = fingerprint
+    status_data["resume_from_phase"] = None
+    if session_id:
+        sessions = status_data.get("sessions", [])
+        if session_id not in sessions:
+            sessions.append(session_id)
+        status_data["sessions"] = sessions
+        status_data["last_session_id"] = session_id
+
+
 def set_context_overflow_continuation_metadata(status_data, args, session_id):
     """Record continuation metadata for a context-overflow outcome in status.json."""
     status_data["context_overflow_count"] = status_data.get("context_overflow_count", 0) + 1
@@ -460,6 +487,18 @@ def action_update(args, refactor_list_path, state_dir):
 
     # Track what status we write to refactor-list.json
     new_status = current_list_status
+    if args.active_dev_branch:
+        rs["active_dev_branch"] = args.active_dev_branch
+    if args.base_branch:
+        rs["base_branch"] = args.base_branch
+    if args.last_fatal_error_code:
+        rs["last_fatal_error_code"] = args.last_fatal_error_code
+    if args.continuation_summary_path:
+        rs["continuation_summary_path"] = args.continuation_summary_path
+    fingerprint = _load_progress_fingerprint(args.progress_fingerprint)
+    if fingerprint is not None:
+        rs["last_progress_fingerprint"] = fingerprint
+
 
     if session_status == "success":
         rs["infra_error_count"] = 0
@@ -503,10 +542,14 @@ def action_update(args, refactor_list_path, state_dir):
 
         if infra_error_count >= max_infra_retries:
             new_status = "failed"
+            rs["continuation_pending"] = False
+            rs["continuation_reason"] = session_status
+            rs["needs_attention"] = True
             if session_id:
                 rs["last_failed_session_id"] = session_id
         else:
-            new_status = "pending"
+            new_status = "in_progress"
+            _set_ai_error_continuation_metadata(rs, args, session_id, session_status)
 
         err = update_refactor_in_list(refactor_list_path, refactor_id, new_status)
         if err:
@@ -532,27 +575,25 @@ def action_update(args, refactor_list_path, state_dir):
     else:
         rs["retry_count"] = rs.get("retry_count", 0) + 1
 
-        cleaned = cleanup_refactor_artifacts(
-            state_dir=state_dir,
-            refactor_id=refactor_id,
-            project_root=args.project_root,
-        )
-
         if rs["retry_count"] >= max_retries:
             new_status = "failed"
+            rs["continuation_pending"] = False
+            rs["continuation_reason"] = session_status
+            rs["needs_attention"] = True
         else:
-            new_status = "pending"
+            new_status = "in_progress"
+            _set_ai_error_continuation_metadata(rs, args, session_id, session_status)
 
         rs["resume_from_phase"] = None
-        rs["sessions"] = []
-        rs["last_session_id"] = None
+        if session_id:
+            rs["last_failed_session_id"] = session_id
 
         err = update_refactor_in_list(refactor_list_path, refactor_id, new_status)
         if err:
             error_out("Failed to update .prizmkit/plans/refactor-list.json: {}".format(err))
             return
 
-    if session_status == "success" and session_id:
+    if session_id:
         sessions = rs.get("sessions", [])
         if session_id not in sessions:
             sessions.append(session_id)
@@ -594,8 +635,9 @@ def action_update(args, refactor_list_path, state_dir):
         summary["degraded_reason"] = session_status
         summary["restart_policy"] = "finalization_retry"
     elif session_status == "infra_error":
-        summary["restart_policy"] = "infra_retry"
+        summary["restart_policy"] = "in_process_ai_error_continuation" if new_status == "in_progress" else "needs_attention"
         summary["infra_error_count"] = rs.get("infra_error_count", 0)
+        summary.update(continuation_metadata_summary(rs))
         summary["artifacts_preserved"] = True
     elif session_status == "context_overflow":
         summary["restart_policy"] = "context_overflow_continuation"
@@ -606,8 +648,9 @@ def action_update(args, refactor_list_path, state_dir):
         summary.update(continuation_metadata_summary(rs))
         summary["artifacts_preserved"] = True
     elif session_status != "success":
-        summary["restart_policy"] = "full_restart"
-        summary["cleanup_performed"] = cleaned
+        summary["restart_policy"] = "in_process_ai_error_continuation" if new_status == "in_progress" else "needs_attention"
+        summary.update(continuation_metadata_summary(rs))
+        summary["artifacts_preserved"] = True
 
     print(json.dumps(summary, indent=2, ensure_ascii=False))
 

package/bundled/dev-pipeline/scripts/utils.py

@@ -745,7 +745,7 @@ def resolve_prompt_platform(project_root):
     if resolution.platform:
         return resolution.platform
     raise RuntimeError(
-        "PrizmKit agents not found. None of .claude/agents/, .codex/agents/, or .codebuddy/agents/ exists. "
+        "PrizmKit platform skills not found. None of .claude/commands/, .agents/skills/, or .codebuddy/skills/ exists. "
         "Run `npx prizmkit install` first, set .prizmkit/config.json ai_cli/platform, or set PRIZMKIT_PLATFORM=claude|codex|codebuddy explicitly."
     )
 

package/bundled/dev-pipeline/templates/bootstrap-prompt.md

@@ -1,8 +1,8 @@
 # Dev-Pipeline Fallback Bootstrap Prompt
 
 > **Note**: This is an emergency fallback template. Normally, `generate-bootstrap-prompt.py`
-> selects a tier-specific template (`bootstrap-tier1.md`, `bootstrap-tier2.md`, or `bootstrap-tier3.md`).
-> If you are seeing this prompt, the tier-specific templates were not found.
+> renders modular full guidance from `templates/sections/`. If sections are unavailable,
+> it tries `bootstrap-tier3.md` before this last-resort fallback.
 
 ## Feature Context
 

package/bundled/dev-pipeline/templates/bootstrap-tier3.md

@@ -50,7 +50,7 @@ You are running in **headless non-interactive mode** with a FINITE context windo
 
 0. **NON-INTERACTIVE MODE** — There is NO human on the other end. NEVER ask for user confirmation, NEVER wait for user input, NEVER use interactive prompts (e.g. "Would you like me to…"). If a skill has an interactive step (e.g. offer remediation, ask for approval), skip it and proceed autonomously. Make decisions based on the data available and move forward.
 
-1. **context-snapshot.md is your single source of truth** — After Phase 1-2 builds it, ALWAYS read context-snapshot.md instead of re-reading individual source files
+1. **context-snapshot.md is your single source of truth** — After Phase 1-2 makes it available, use it instead of re-reading individual source files. If you just created or updated it in this session, continue from the in-context content; read it only when resuming/continuing, after context compaction/loss, or when handing work to a subagent.
 2. **Never re-read your own writes** — After you create/modify a file, do NOT read it back to verify. Trust your write was correct.
 3. **Never re-read unmodified files** — Once you have read a file into context, do NOT read it again unless it was modified by an Edit/Write operation after your last read. Use platform file/search tools for targeted lookup; when shell-free helper lookup is needed, run `{{RUNTIME_HELPER_CMD}} text search <file> --pattern "<pattern>" --json`, then read only the smallest bounded range with the platform file-reading tool instead of full-file Read. Before issuing a Read, ask yourself: "Have I read this file before, and has it changed since?" If the answer is "no, it hasn't changed" — do NOT read it. This applies to all files including those summarized in context-snapshot.md Section 4. Reviewer subagents must receive context-snapshot.md and rely on it rather than independently re-reading the same files the orchestrator already read.
 4. **Stay focused** — Do NOT explore code unrelated to this feature. No curiosity-driven reads.
@@ -205,14 +205,13 @@ Before starting, check plan.md Tasks:
 
 **Build artifacts rule**: After any build/compile command, ensure output is in `.gitignore`. Never commit binaries/build output.
 
-**3a.** Run `/prizmkit-implement` directly — you (the orchestrator) execute the full cycle:
-- Read plan.md Tasks from `.prizmkit/specs/{{FEATURE_SLUG}}/`
-- Use context-snapshot.md Section 4 File Manifest for targeted reads — do NOT re-read files already summarized there
+**3a.** Run `/prizmkit-implement` directly:
+- Continue from the `spec.md` / `plan.md` / `context-snapshot.md` content you created in Phase 1-2; do not re-read self-authored artifacts just because implementation starts
+- Re-read `plan.md` or `context-snapshot.md` only when resuming/continuing, when current context was compacted or lost, or when task checkbox state may have changed outside this session
+- If implementation is delegated to a Dev subagent, the subagent must read `spec.md` if the artifact exists, plus `plan.md` and `context-snapshot.md`, before editing
+- Use context-snapshot.md Section 4 File Manifest for targeted source reads; avoid re-reading source files already summarized there unless they changed after the snapshot
 - Implements task-by-task, marking each `[x]` immediately
-- Creates/updates L2 `.prizm` docs when creating new modules
-- Defers scoped/full test execution until after the code-review loop completes
-- Preserves Reviewer behavior; only the top-level implementation Dev handoff is removed
-- If plan.md has >5 tasks: update checkpoints after every 3 tasks
+- Remeber to update checkpoints after finishing a series of tasks
 
 **3b. Focused checks only (no test gate here)**:
 - During implementation, do not run mandatory periodic or full-suite tests.
@@ -241,7 +240,7 @@ For each Verification Gate from the Task Contract, write one evidence line to Im
 
 Review the completed implementation before the scoped feature test gate runs. Use `artifact_dir=.prizmkit/specs/{{FEATURE_SLUG}}/` so the code-review skill can read spec, plan, context snapshot, implementation log, and current diff artifacts. Do not require or read a prior `/prizmkit-test` report here; test generation and test-fix looping happen in the next phase.
 
-Run `/prizmkit-code-review` with `artifact_dir=.prizmkit/specs/{{FEATURE_SLUG}}/`. The code-review skill must launch Reviewer in `current-workspace / active-checkout / no-worktree` mode so it reviews the active checkout rather than an isolated copy; do not request worktree isolation or use `.claude/worktrees/...` / `.prizmkit/state/worktrees/...`. Do not set `USE_WORKTREE` for reviewer launches.
+Run `/prizmkit-code-review` with `artifact_dir=.prizmkit/specs/{{FEATURE_SLUG}}/`. The skill chooses a platform-supported independent Reviewer strategy and verifies that the Reviewer's snapshot is equivalent to the current workspace under its skill-owned workspace protocol. Do not prescribe Agent launch parameters from this bootstrap prompt.
 
 The skill runs an internal review-fix loop (Reviewer Agent → filter → orchestrator fix, max 3 rounds) and writes `review-report.md` to the artifact directory.
 
@@ -254,7 +253,7 @@ If GATE:MISSING:
 - Do not re-run `/prizmkit-code-review` in an unbounded report-repair loop.
 - Perform one bounded status check; retry at most once: inspect the skill output, `review-report.md` path, and any Reviewer agent spawn messages.
 - If the missing report is caused by Reviewer agent spawn or review skill error, retry `/prizmkit-code-review` at most once only if it appears transient.
-- If the report is still missing after that single check/retry, write `failure-log.md` with the spawn/skill error and last observable state, then either perform a safe inline fallback review (spec/plan/diff/tests → write `review-report.md` with `## Verdict`) or stop with a clear recovery failure.
+- If the report is still missing after that single check/retry, write `failure-log.md` with the spawn/skill error and last observable state, then stop with a clear recovery failure. Do not synthesize a passing report through Main-Agent self-review.
 
 Read `review-report.md` and check the Verdict:
 - `PASS` → proceed to next phase
@@ -507,7 +506,6 @@ Rules for writing completion notes:
 |----------|------|
 | Feature Artifacts Dir | `.prizmkit/specs/{{FEATURE_SLUG}}/` |
 | Context Snapshot | `.prizmkit/specs/{{FEATURE_SLUG}}/context-snapshot.md` |
-| Team Config | `{{TEAM_CONFIG_PATH}}` |
 | Project Root | {{PROJECT_ROOT}} |
 | Feature List Path | {{FEATURE_LIST_PATH}} |
 

package/bundled/dev-pipeline/templates/bugfix-bootstrap-prompt.md

@@ -180,6 +180,7 @@ After implement completes, verify:
 If `FAST_PATH=true` (≤ 2 tasks, obvious root cause), skip this phase entirely and mark checkpoint step `prizmkit-code-review` as `skipped` before continuing.
 
 Run `/prizmkit-code-review` with `artifact_dir=.prizmkit/bugfix/{{BUG_ID}}/`:
+- The skill chooses a platform-supported independent Reviewer strategy and verifies a workspace-equivalent snapshot under its skill-owned protocol
 - The skill runs an internal review-fix loop (Reviewer → filter → orchestrator fix, max 3 rounds) and writes `review-report.md`
 - `review-report.md` must contain `## Verdict`
 
@@ -191,7 +192,7 @@ After `/prizmkit-code-review` returns, verify the review report:
 If GATE:MISSING:
 - Do not re-run `/prizmkit-code-review` in an unbounded report-repair loop.
 - Perform one bounded status check; retry `/prizmkit-code-review` at most once only if the missing report appears caused by a transient team/config/lock error.
-- If the report is still missing after that single check/retry, write `failure-log.md` with the spawn/skill error and last observable state, then either perform a safe inline fallback review (spec/plan/diff/tests → write `review-report.md` with `## Verdict`) or stop with a clear recovery failure.
+- If the report is still missing after that single check/retry, write `failure-log.md` with the spawn/skill error and last observable state, then stop with a clear recovery failure. Do not replace the independent Reviewer with Main-Agent self-review.
 
 Read `review-report.md` and check the Verdict:
 - `PASS` → proceed

package/bundled/dev-pipeline/templates/refactor-bootstrap-prompt.md

@@ -139,7 +139,7 @@ Apply the browser behavior preservation protocol where UI behavior can be affect
 
 ### Phase 4: Review — Code Review & Behavior Verification
 
-Run `/prizmkit-code-review` with `artifact_dir=.prizmkit/refactor/{{REFACTOR_ID}}/` directly from the main orchestrator.
+Run `/prizmkit-code-review` with `artifact_dir=.prizmkit/refactor/{{REFACTOR_ID}}/` directly from the main orchestrator. The skill loads its skill-owned `references/reviewer-agent-prompt.md` and `references/reviewer-workspace-protocol.md`, chooses a platform-supported independent Reviewer strategy, and verifies a workspace-equivalent snapshot.
 
 Review requirements:
 - verify refactor quality against spec.md and plan.md;
@@ -147,7 +147,7 @@ Review requirements:
 - run or confirm relevant tests;
 - write `.prizmkit/refactor/{{REFACTOR_ID}}/review-report.md` with `## Verdict`.
 
-If the report is missing after one bounded status check, write `failure-log.md` and either perform a safe inline fallback review or stop with a clear recovery failure.
+If the report is missing after one bounded status check, write `failure-log.md` and stop with a clear recovery failure. Do not replace the independent Reviewer with Main-Agent self-review.
 
 **Checkpoint update**: set step `prizmkit-code-review` to `completed` in `{{CHECKPOINT_PATH}}`.
 

package/bundled/dev-pipeline/templates/sections/bugfix-phase-review.md

@@ -12,9 +12,9 @@ If `FAST_PATH=true`, skip this phase intentionally:
 
 Then continue to the next checkpoint step.
 
-Otherwise, run `/prizmkit-code-review` with `artifact_dir=.prizmkit/bugfix/{{BUG_ID}}/`.
+Otherwise, run `/prizmkit-code-review` with `artifact_dir=.prizmkit/bugfix/{{BUG_ID}}/`. The skill loads its skill-owned Reviewer prompt and workspace protocol, chooses a platform-supported subagent strategy, and verifies that the Reviewer's snapshot is equivalent to the current workspace before review.
 
-The skill runs its internal Reviewer Agent → filter → orchestrator fix loop and writes `review-report.md`.
+The skill runs its internal Reviewer Agent → filter → orchestrator fix loop and writes `review-report.md`. Do not prescribe Agent launch parameters from this pipeline template; the review skill owns strategy selection and workspace-equivalence verification.
 
 **Gate Check — Review Report**:
 
@@ -24,6 +24,6 @@ The skill runs its internal Reviewer Agent → filter → orchestrator fix loop
 
 - `PASS` → proceed.
 - `NEEDS_FIXES` → the skill exhausted its internal rounds; log remaining findings and proceed only when they do not block acceptance criteria.
-- Missing report → one bounded retry or safe inline fallback review, then write `failure-log.md` if still blocked.
+- Missing report → one bounded retry, then write `failure-log.md` and stop if the independent review protocol remains blocked; do not replace it with Main-Agent self-review.
 
 **Checkpoint update**: Set step `prizmkit-code-review` to `completed` when review is complete, or `skipped` for FAST_PATH.

package/bundled/dev-pipeline/templates/sections/context-budget-rules.md

@@ -4,7 +4,7 @@ You are running in **headless non-interactive mode** with a FINITE context windo
 
 0. **NON-INTERACTIVE MODE** — There is NO human on the other end. NEVER ask for user confirmation, NEVER wait for user input, NEVER use interactive prompts (e.g. "Would you like me to…"). If a skill has an interactive step (e.g. offer remediation, ask for approval), skip it and proceed autonomously. Make decisions based on the data available and move forward.
 
-1. **context-snapshot.md is your single source of truth** — After it is built, ALWAYS read context-snapshot.md instead of re-reading individual source files
+1. **context-snapshot.md is your single source of truth** — After it is available, use it instead of re-reading individual source files. If you just created or updated it in this session, continue from the in-context content; read it only when resuming/continuing, after context compaction/loss, or when handing work to a subagent.
 2. **Never re-read your own writes** — After you create/modify a file, do NOT read it back to verify. Trust your write was correct.
 3. **Never re-read unmodified files** — Once you have read a file into context, do NOT read it again unless it was modified by an Edit/Write operation after your last read. Use platform file/search tools for targeted lookup; when shell-free helper lookup is needed, run `{{RUNTIME_HELPER_CMD}} text search <file> --pattern "<pattern>" --json`, then read only the smallest bounded range with the platform file-reading tool instead of full-file Read. Before issuing a Read, ask yourself: "Have I read this file before, and has it changed since?" If the answer is "no, it hasn't changed" — do NOT read it. This applies to all files including those summarized in context-snapshot.md Section 4. Reviewer subagents must receive context-snapshot.md and rely on it rather than independently re-reading the same files the orchestrator already read.
 4. **Stay focused** — Do NOT explore code unrelated to this feature. No curiosity-driven reads.

package/bundled/dev-pipeline/templates/sections/critical-paths-agent.md

@@ -4,5 +4,4 @@
 |----------|------|
 | Feature Artifacts Dir | `.prizmkit/specs/{{FEATURE_SLUG}}/` |
 | Context Snapshot | `.prizmkit/specs/{{FEATURE_SLUG}}/context-snapshot.md` |
-| Reviewer Agent Def | {{REVIEWER_SUBAGENT_PATH}} |
 | Project Root | {{PROJECT_ROOT}} |

package/bundled/dev-pipeline/templates/sections/critical-paths-full.md

@@ -4,7 +4,5 @@
 |----------|------|
 | Feature Artifacts Dir | `.prizmkit/specs/{{FEATURE_SLUG}}/` |
 | Context Snapshot | `.prizmkit/specs/{{FEATURE_SLUG}}/context-snapshot.md` |
-| Team Config | `{{TEAM_CONFIG_PATH}}` |
-| Reviewer Agent Def | {{REVIEWER_SUBAGENT_PATH}} |
 | Project Root | {{PROJECT_ROOT}} |
 | Feature List Path | {{FEATURE_LIST_PATH}} |

package/bundled/dev-pipeline/templates/sections/phase-commit-full.md

@@ -1,12 +1,17 @@
 ### Architecture Sync & Commit (SINGLE COMMIT) — DO NOT SKIP
 
-**a.** Check if feature already committed:
+**a.** Check if feature already committed by inspecting recent git history for `{{FEATURE_ID}}`:
 ```bash
 git log --oneline
 ```
 - If a commit for `{{FEATURE_ID}}` already exists → **skip d** (do NOT run /prizmkit-committer, do NOT run git reset, do NOT stage or unstage anything). Proceed directly to e Final verification.
 - If no existing commit → proceed normally with b–d.
 
+**Feature Documentation Policy**:
+- Run `/prizmkit-retrospective` before commit to update `.prizmkit/prizm-docs/` with durable structural knowledge from this feature.
+- Capture architecture knowledge when the feature adds interfaces, dependencies, observable behavior, modules, or durable TRAPS/RULES/DECISIONS.
+- Commit prefix: `feat({{FEATURE_ID}}):`.
+
 **b.** Run `/prizmkit-retrospective` (**before commit**, maintains `.prizmkit/prizm-docs/` architecture index):
 1. **Structural sync**: update KEY_FILES/INTERFACES/DEPENDENCIES/file counts for changed modules
 2. **Architecture knowledge**: extract TRAPS, RULES, DECISIONS from completed work into `.prizmkit/prizm-docs/`
@@ -32,6 +37,8 @@ git status --short
 ```
 Working tree MUST be clean after this step. If any feature-related files remain, stage them explicitly and amend the SAME commit with `git commit --amend --no-edit`; do NOT create a separate commit.
 
+**Exception**: `session-summary.md` in the artifact directory is a local cross-session artifact generated by `/prizmkit-committer` — it is NOT committed to git. Ignore it in the clean-tree check.
+
 **f.** Write completion summary for downstream dependency context:
 
 Write `.prizmkit/specs/{{FEATURE_SLUG}}/completion-summary.json` with the key changes from this session. This file is NOT committed to git — the pipeline runner reads it to propagate context to dependent features.

package/bundled/dev-pipeline/templates/sections/phase-implement-agent.md

@@ -11,14 +11,13 @@ Before starting, check plan.md Tasks:
 
 **Build artifacts rule**: After any build/compile command, ensure output is in `.gitignore`. Never commit binaries/build output.
 
-**3a.** Run `/prizmkit-implement` directly — you (the orchestrator) execute the full cycle:
-- Read plan.md Tasks from `.prizmkit/specs/{{FEATURE_SLUG}}/`
-- Use context-snapshot.md Section 4 File Manifest for targeted reads — do NOT re-read files already summarized there
+**3a.** Run `/prizmkit-implement` directly:
+- Continue from the `spec.md` / `plan.md` / `context-snapshot.md` content you created in Phase 1-2; do not re-read self-authored artifacts just because implementation starts
+- Re-read `plan.md` or `context-snapshot.md` only when resuming/continuing, when current context was compacted or lost, or when task checkbox state may have changed outside this session
+- If implementation is delegated to a Dev subagent, the subagent must read `spec.md` if the artifact exists, plus `plan.md` and `context-snapshot.md`, before editing
+- Use context-snapshot.md Section 4 File Manifest for targeted source reads; avoid re-reading source files already summarized there unless they changed after the snapshot
 - Implements task-by-task, marking each `[x]` immediately
-- Creates/updates L2 `.prizm` docs when creating new modules
-- Defers scoped/full test execution until after the code-review loop completes
-- Preserves Reviewer behavior; only the top-level implementation Dev handoff is removed
-- If plan.md has >5 tasks: update checkpoints after every 3 tasks
+- Remeber to update checkpoints after finishing a series of tasks
 
 **3b. Focused checks only (no test gate here)**:
 - During implementation, do not run mandatory periodic or full-suite tests.

package/bundled/dev-pipeline/templates/sections/phase-implement-full.md

@@ -11,14 +11,13 @@ Before starting, check plan.md Tasks:
 
 **Build artifacts rule**: After any build/compile command, ensure output is in `.gitignore`. Never commit binaries/build output.
 
-**3a.** Run `/prizmkit-implement` directly — you (the orchestrator) execute the full cycle:
-- Read plan.md Tasks from `.prizmkit/specs/{{FEATURE_SLUG}}/`
-- Use context-snapshot.md Section 4 File Manifest for targeted reads — do NOT re-read files already summarized there
+**3a.** Run `/prizmkit-implement` directly:
+- Continue from the `spec.md` / `plan.md` / `context-snapshot.md` content you created in Phase 1-2; do not re-read self-authored artifacts just because implementation starts
+- Re-read `plan.md` or `context-snapshot.md` only when resuming/continuing, when current context was compacted or lost, or when task checkbox state may have changed outside this session
+- If implementation is delegated to a Dev subagent, the subagent must read `spec.md` if the artifact exists, plus `plan.md` and `context-snapshot.md`, before editing
+- Use context-snapshot.md Section 4 File Manifest for targeted source reads; avoid re-reading source files already summarized there unless they changed after the snapshot
 - Implements task-by-task, marking each `[x]` immediately
-- Creates/updates L2 `.prizm` docs when creating new modules
-- Defers scoped/full test execution until after the code-review loop completes
-- Preserves Reviewer behavior; only the top-level implementation Dev handoff is removed
-- If plan.md has >5 tasks: update checkpoints after every 3 tasks
+- Remeber to update checkpoints after finishing a series of tasks
 
 **3b. Focused checks only (no test gate here)**:
 - During implementation, do not run mandatory periodic or full-suite tests.

package/bundled/dev-pipeline/templates/sections/phase-implement-lite.md

@@ -6,8 +6,10 @@
 - Follow Context Budget Rules §10 after build/compile commands.
 
 **3a.** Run `/prizmkit-implement` — this handles the implementation cycle:
-- Reads plan.md Tasks section from `.prizmkit/specs/{{FEATURE_SLUG}}/`
-- Reads context from `context-snapshot.md` (Prizm docs, TRAPS, file manifest)
+- Continue from the plan/context you just created; do not re-read self-authored artifacts just because implementation starts
+- Re-read plan.md or `context-snapshot.md` only when resuming/continuing, when current context was compacted or lost, or when task checkbox state may have changed outside this session
+- If implementation is delegated to a subagent, the subagent must read `plan.md` and `context-snapshot.md` before editing
+- Uses context from `context-snapshot.md` (Prizm docs, TRAPS, file manifest) for targeted source reads; avoid re-reading source files already summarized there unless they changed after the snapshot
 - Implements task-by-task, marking each `[x]` immediately
 - Creates/updates L2 `.prizm` docs when creating new modules or significantly modifying existing ones — AI selectively decides which modules warrant L2 based on complexity and importance
 - Does not run the scoped/full test gate; testing happens after code review in the PrizmKit Test phase