Fix Claude review stream parsing and clean verdict findings

.claude/settings.json

@@ -23,11 +23,11 @@
         ]
       },
       {
-        "matcher": "TaskCreate|TaskUpdate",
+        "matcher": "Skill",
         "hooks": [
           {
             "type": "command",
-            "command": "sh -c 'if ! command -v entire >/dev/null 2>&1; then exit 0; fi; exec entire hooks claude-code post-todo'"
+            "command": "sh -c 'if ! command -v entire >/dev/null 2>&1; then exit 0; fi; exec entire hooks claude-code post-task'"
           }
         ]
       }
@@ -80,6 +80,28 @@
         ]
       }
     ],
+    "StopFailure": [
+      {
+        "matcher": "",
+        "hooks": [
+          {
+            "type": "command",
+            "command": "sh -c 'if ! command -v entire >/dev/null 2>&1; then exit 0; fi; exec entire hooks claude-code stop-failure'"
+          }
+        ]
+      }
+    ],
+    "SubagentStart": [
+      {
+        "matcher": "workflow-subagent",
+        "hooks": [
+          {
+            "type": "command",
+            "command": "sh -c 'if ! command -v entire >/dev/null 2>&1; then exit 0; fi; exec entire hooks claude-code subagent-start'"
+          }
+        ]
+      }
+    ],
     "SubagentStop": [
       {
         "matcher": "",

.claude/skills/agent-integration/SKILL.md

@@ -19,7 +19,7 @@ Collect these before starting (ask the user if not provided):
 
 | Parameter | Description | How to derive |
 |-----------|-------------|---------------|
-| `AGENT_NAME` | Human-readable name (e.g., "Gemini CLI") | User provides |
+| `AGENT_NAME` | Human-readable name (e.g., "Copilot CLI") | User provides |
 | `AGENT_PACKAGE` | Go package dir name — **no hyphens** | Lowercase, remove hyphens/spaces |
 | `AGENT_KEY` | Registry key for `agent.Register()` and `entire enable` | Check existing patterns in `cmd/entire/cli/agent/registry.go` |
 | `AGENT_SLUG` | Filesystem/URL-safe slug (kebab-case) used in E2E runner filenames and script names | Kebab-case of agent name; align with existing entries in `e2e/agents/` |

.claude/skills/agent-integration/test-writer.md

@@ -20,7 +20,7 @@ Read these files to understand the existing test patterns.
 3. `e2e/agents/agent.go` — `Agent` interface (`Name`, `Binary`, `EntireAgent`, `PromptPattern`, `TimeoutMultiplier`, `RunPrompt`, `StartSession`, `Bootstrap`, `IsTransientError`), `Register()` for agent self-registration in `init()`, `RegisterGate()` for concurrency limits, `AcquireSlot`/`ReleaseSlot` for gating
 4. `e2e/agents/tmux.go` — `TmuxSession` for interactive PTY-based tests: `NewTmuxSession`, `Send`, `SendKeys`, `WaitFor` (with settle-time logic), `Capture`, `Close`
 5. `e2e/testutil/assertions.go` — Rich assertion helpers: `AssertFileExists`, `WaitForFileExists`, `AssertNewCommits`, `WaitForCheckpoint`, `AssertCheckpointAdvanced`, `AssertHasCheckpointTrailer`, `AssertCheckpointExists`, `AssertCommitLinkedToCheckpoint`, `AssertCheckpointMetadataComplete`, `ValidateCheckpointDeep`, and many more
-6. `e2e/testutil/metadata.go` — `CheckpointMetadata`, `SessionMetadata`, `TokenUsage`, `Attribution`, `SessionRef` types; `CheckpointPath()` helper for sharded directory layout
+6. `e2e/testutil/metadata.go` — `CheckpointMetadata`, `SessionMetadata`, `TokenUsage`, `SessionRef` types; `CheckpointPath()` helper for sharded directory layout
 7. `e2e/entire/entire.go` — CLI wrapper: `BinPath()` (builds from source or uses `E2E_ENTIRE_BIN`), `Enable`, `Disable`, `Doctor`, `CleanDryRun`, `CleanForce`, `Explain`, `AttachWithEnv`, `Resume` / `ResumeWithEnv`
 8. `e2e/testutil/artifacts.go` — Automatic artifact capture via `t.Cleanup`: `CaptureArtifacts` saves git-log, git-tree, checkpoint metadata, entire logs, and tmux pane content
 
@@ -49,7 +49,7 @@ Read `cmd/entire/cli/agent/$AGENT_PACKAGE/AGENT.md` (the one-pager from the rese
 
 Add a new `Agent` implementation in `e2e/agents/${agent_slug}.go`:
 
-**Pattern to follow** (based on existing implementations like `claude.go`, `gemini.go`, `opencode.go`):
+**Pattern to follow** (based on existing implementations like `claude.go`, `codex.go`, `opencode.go`):
 
 ```go
 package agents
@@ -153,7 +153,7 @@ func (a *${AgentName}) StartSession(ctx context.Context, dir string) (Session, e
 
 Key implementation details:
 - Self-register in `init()` with `Register()`, gated by `E2E_AGENT` env var
-- Use `RegisterGate("name", N)` if the agent's API has strict rate limits (e.g., Gemini uses gate of 1)
+- Use `RegisterGate("name", N)` if the agent's API has strict rate limits (e.g., Factory AI Droid uses gate of 1)
 - `Bootstrap()` handles CI-specific one-time setup (auth config, API key injection)
 - `IsTransientError()` identifies retryable API failures — `RepoState.RunPrompt` retries once on transient errors
 - `RunPrompt()` uses `exec.CommandContext` with `Setpgid: true` and process-group kill for clean cancellation
@@ -199,7 +199,7 @@ Use `/commit` to commit all files.
 - **Interactive tests**: Use `s.StartSession`, `s.Send`, `s.WaitFor` — tmux pane is auto-captured in artifacts
 - **Run commands**: `mise run test:e2e --agent ${slug} TestName` — see `e2e/README.md` for all options
 - **E2E tests are run during the implement phase**: This phase only creates the runner. The implement phase runs E2E tests at each tier to drive development.
-- **Debugging failures**: If tests fail during the implement phase, use `/e2e:debug` with the artifact directory to diagnose CLI-level issues (hooks, checkpoints, session phases, attribution)
+- **Debugging failures**: If tests fail during the implement phase, use `/e2e:debug` with the artifact directory to diagnose CLI-level issues (hooks, checkpoints, session phases, file tracking)
 
 ## Output
 

.claude/skills/e2e/debug.md

@@ -68,7 +68,7 @@ Cross-reference console.log (what happened) with the test (what should have happ
 | Wrong checkpoint content | Check `git-tree.txt` for checkpoint branch files, `checkpoint-metadata/` for session info |
 | Hooks didn't fire | Check `entire-logs/entire.log` for missing hook entries (session-start, user-prompt-submit, stop, post-commit) |
 | Stash/unstash problems | Check `entire-logs/entire.log` for stash-related log lines, `git-log.txt` for commit ordering |
-| Attribution issues | Check `checkpoint-metadata/` for `files_touched`, session metadata for attribution data |
+| File-tracking issues | Check `checkpoint-metadata/` for `files_touched` and the session metadata's `files_touched` |
 | Strategy mismatch | Check `entire-logs/entire.log` for `strategy` field, verify auto-commit vs manual-commit behavior |
 
 ### 5. Deep dive files
@@ -84,5 +84,5 @@ Identify whether the issue is in:
 - **CLI hooks** (prepare-commit-msg, commit-msg, post-commit)
 - **Session management** (phase transitions, session tracking)
 - **Checkpoint creation** (branch management, metadata writing)
-- **Attribution** (file tracking, prompt correlation)
+- **File tracking** (files touched, prompt correlation)
 - **Strategy logic** (auto-commit vs manual-commit behavior)

.claude/skills/test-repo/SKILL.md

@@ -82,8 +82,7 @@ Execute these steps in order:
 ```bash
 .claude/skills/test-repo/test-harness.sh verify-commit
 .claude/skills/test-repo/test-harness.sh verify-session-state
-.claude/skills/test-repo/test-harness.sh verify-shadow-branch
-.claude/skills/test-repo/test-harness.sh verify-metadata-branch
+.claude/skills/test-repo/test-harness.sh verify-no-shadow-branch
 .claude/skills/test-repo/test-harness.sh list-pending-checkpoints
 ```
 
@@ -92,10 +91,10 @@ Expected results:
 | Check | Result |
 |-------|--------|
 | Active branch | Optional Entire-Checkpoint: trailer |
-| Session state | ✓ Exists |
-| Shadow branch | ✓ entire/{hash} |
-| Metadata branch | ✓ entire/checkpoints/v1 |
-| Pending checkpoints | ✓ At least 1 |
+| Session state | ✓ Exists, with the turn's files in `files_touched` |
+| Shadow branch | ✓ None (turn ends write nothing to git) |
+| Pending checkpoints | ✓ A next-checkpoint preview (`is_next_checkpoint: true`) |
+| Metadata branch | Written only when you commit (see "Test User Commits") |
 
 #### 4. Check Listing After Further Changes
 
@@ -105,9 +104,9 @@ Expected results:
 ```
 
 **Expected Behavior:**
-- The checkpoint from step 2 is still listed; the new uncommitted changes do not
-  disturb it. There is no restore step: the CLI has no `rewind` command, and
-  nothing writes a checkpoint back over the worktree.
+- The next-checkpoint preview from step 3 is still listed; the new
+  uncommitted changes do not disturb it. There is no restore step: the CLI has
+  no `rewind` command, and nothing writes a checkpoint back over the worktree.
 
 #### 5. Cleanup
 
@@ -131,16 +130,16 @@ go build -o /tmp/entire-bin ./cmd/entire && \
 .claude/skills/test-repo/test-harness.sh create-files && \
 .claude/skills/test-repo/test-harness.sh create-transcript && \
 .claude/skills/test-repo/test-harness.sh stop-session && \
-.claude/skills/test-repo/test-harness.sh verify-metadata-branch && \
+.claude/skills/test-repo/test-harness.sh verify-no-shadow-branch && \
 .claude/skills/test-repo/test-harness.sh list-pending-checkpoints
 ```
 
 ## Expected Results by Strategy
 
 ### Manual-Commit Strategy (default)
 - Active branch commits: **NO modifications** (no commits created by Entire)
-- Shadow branches: `entire/<commit-hash[:7]>` created for checkpoints
-- Metadata: stored on both shadow branches and `entire/checkpoints/v1` branch (condensed on user commits)
+- Turn ends: recorded in session state only (`.git/entire-sessions/`); no shadow branch is written
+- Metadata: written to `entire/checkpoints/v1` (or per-checkpoint refs) when you commit
 - AllowsMainBranch: **true** (safe on main/master)
 
 ## Additional Testing (Optional)
@@ -167,8 +166,8 @@ echo "{\"session_id\": \"$SESSION_ID\", \"transcript_path\": \"$TRANSCRIPT_DIR/t
   ENTIRE_TEST_CLAUDE_PROJECT_DIR="$TRANSCRIPT_DIR" \
   /tmp/entire-bin hooks claude-code post-task
 
-# Verify task checkpoint created
-/tmp/entire-bin checkpoint list --pending --json | jq '.[] | select(.is_task_checkpoint == true)'
+# Verify the task record is pending (listed in this session's next-checkpoint preview)
+/tmp/entire-bin checkpoint list --pending --json | jq '.[] | select(.is_next_checkpoint == true) | .next_checkpoint.task_records'
 ```
 
 ### Test User Commits (Condensation)
@@ -182,9 +181,6 @@ git commit -m "Add greeting function"
 
 # Verify logs condensed to entire/checkpoints/v1
 git show entire/checkpoints/v1 --stat | grep -E "^[0-9a-f]{2}/[0-9a-f]"
-
-# Verify shadow branch still exists
-git branch -a | grep "entire/[0-9a-f]"
 ```
 
 ## Available Claude Code Hooks
@@ -193,10 +189,10 @@ All hooks use the command: `entire hooks claude-code <hook-name>`
 
 - `user-prompt-submit` - Called when user submits a prompt (before session starts)
 - `session-start` - Called when session starts
-- `stop` - Called when session stops (creates checkpoint)
+- `stop` - Called when the agent's turn ends (records the turn in session state; the checkpoint is written at commit)
 - `pre-task` - Called before Task tool execution
 - `post-task` - Called after Task tool execution
-- `post-todo` - Called after TodoWrite tool execution (for incremental checkpoints)
+- `post-todo` - No longer installed; still accepted so configs written by older CLIs keep working, and records nothing
 
 ## Report Format
 

.claude/skills/test-repo/test-harness.sh

@@ -123,14 +123,15 @@ verify-session-state)
   fi
   ;;
 
-verify-shadow-branch)
-  echo "==> Verifying shadow branch..."
+verify-no-shadow-branch)
+  echo "==> Verifying no shadow branch was written..."
   cd "$REPO_DIR"
 
-  if git branch -a | grep -E "entire/[0-9a-f]"; then
-    echo "✓ Shadow branch exists"
+  if git branch -a | grep -E "entire/[0-9a-f]{7,}(-[0-9a-f]{6})?$"; then
+    echo "✗ Shadow branch found: turn ends must not write one"
+    exit 1
   else
-    echo "Note: No shadow branch"
+    echo "✓ No shadow branch"
   fi
   ;;
 
@@ -192,7 +193,7 @@ info)
   echo "  stop-session             - Stop session (create checkpoint)"
   echo "  verify-commit            - Verify active branch commit"
   echo "  verify-session-state     - Verify session state files"
-  echo "  verify-shadow-branch     - Verify shadow branch exists"
+  echo "  verify-no-shadow-branch  - Verify no shadow branch was written"
   echo "  verify-metadata-branch   - Verify metadata branch exists"
   echo "  list-pending-checkpoints - List pending (not yet condensed) checkpoints"
   echo "  create-changes           - Create changes on top of the checkpoint"

.entire/settings.json

@@ -4,8 +4,7 @@
     "checkpoint_remote": {
       "provider": "github",
       "repo": "entireio/cli-checkpoints"
-    },
-    "filtered_fetches": true
+    }
   },
   "strategy": "manual-commit",
   "checkpoints": {

.gemini/settings.json

@@ -3,147 +3,5 @@
     "fileName": [
       "AGENTS.md"
     ]
-  },
-  "hooks": {
-    "AfterAgent": [
-      {
-        "hooks": [
-          {
-            "name": "entire-after-agent",
-            "type": "command",
-            "command": "sh -c 'if ! command -v entire >/dev/null 2>&1; then exit 0; fi; exec entire hooks gemini after-agent'"
-          }
-        ]
-      }
-    ],
-    "AfterModel": [
-      {
-        "hooks": [
-          {
-            "name": "entire-after-model",
-            "type": "command",
-            "command": "sh -c 'if ! command -v entire >/dev/null 2>&1; then exit 0; fi; exec entire hooks gemini after-model'"
-          }
-        ]
-      }
-    ],
-    "AfterTool": [
-      {
-        "matcher": "*",
-        "hooks": [
-          {
-            "name": "entire-after-tool",
-            "type": "command",
-            "command": "sh -c 'if ! command -v entire >/dev/null 2>&1; then exit 0; fi; exec entire hooks gemini after-tool'"
-          }
-        ]
-      }
-    ],
-    "BeforeAgent": [
-      {
-        "hooks": [
-          {
-            "name": "entire-before-agent",
-            "type": "command",
-            "command": "sh -c 'if ! command -v entire >/dev/null 2>&1; then exit 0; fi; exec entire hooks gemini before-agent'"
-          }
-        ]
-      }
-    ],
-    "BeforeModel": [
-      {
-        "hooks": [
-          {
-            "name": "entire-before-model",
-            "type": "command",
-            "command": "sh -c 'if ! command -v entire >/dev/null 2>&1; then exit 0; fi; exec entire hooks gemini before-model'"
-          }
-        ]
-      }
-    ],
-    "BeforeTool": [
-      {
-        "matcher": "*",
-        "hooks": [
-          {
-            "name": "entire-before-tool",
-            "type": "command",
-            "command": "sh -c 'if ! command -v entire >/dev/null 2>&1; then exit 0; fi; exec entire hooks gemini before-tool'"
-          }
-        ]
-      }
-    ],
-    "BeforeToolSelection": [
-      {
-        "hooks": [
-          {
-            "name": "entire-before-tool-selection",
-            "type": "command",
-            "command": "sh -c 'if ! command -v entire >/dev/null 2>&1; then exit 0; fi; exec entire hooks gemini before-tool-selection'"
-          }
-        ]
-      }
-    ],
-    "Notification": [
-      {
-        "hooks": [
-          {
-            "name": "entire-notification",
-            "type": "command",
-            "command": "sh -c 'if ! command -v entire >/dev/null 2>&1; then exit 0; fi; exec entire hooks gemini notification'"
-          }
-        ]
-      }
-    ],
-    "PreCompress": [
-      {
-        "hooks": [
-          {
-            "name": "entire-pre-compress",
-            "type": "command",
-            "command": "sh -c 'if ! command -v entire >/dev/null 2>&1; then exit 0; fi; exec entire hooks gemini pre-compress'"
-          }
-        ]
-      }
-    ],
-    "SessionEnd": [
-      {
-        "matcher": "exit",
-        "hooks": [
-          {
-            "name": "entire-session-end-exit",
-            "type": "command",
-            "command": "sh -c 'if ! command -v entire >/dev/null 2>&1; then exit 0; fi; exec entire hooks gemini session-end'"
-          }
-        ]
-      },
-      {
-        "matcher": "logout",
-        "hooks": [
-          {
-            "name": "entire-session-end-logout",
-            "type": "command",
-            "command": "sh -c 'if ! command -v entire >/dev/null 2>&1; then exit 0; fi; exec entire hooks gemini session-end'"
-          }
-        ]
-      }
-    ],
-    "SessionStart": [
-      {
-        "hooks": [
-          {
-            "name": "entire-session-start",
-            "type": "command",
-            "command": "sh -c 'if ! command -v entire >/dev/null 2>&1; then printf \"%s\\n\" \"{\\\"systemMessage\\\":\\\"Entire CLI is enabled but not installed or not on PATH. Installation guide: https://docs.entire.io/cli/installation#installation-methods\\\"}\"; exit 0; fi; exec entire hooks gemini session-start'"
-          }
-        ]
-      }
-    ]
-  },
-  "hooksConfig": {
-    "enabled": true
-  },
-  "tools": {
-    "enableHooks": true
   }
 }