claude-dev-env 2.12.0 → 2.14.0

package/hooks/hooks_constants/CLAUDE.md

@@ -1,113 +1 @@
-# hooks/hooks_constants
-
-Shared constant modules imported by hooks throughout the `hooks/` tree. Each file holds the tunables for one hook or one cross-cutting concern, keeping magic values out of the hook scripts themselves.
-
-## Files
-
-| File | What it holds |
-|---|---|
-| `__init__.py` | Package marker (`# pragma: no-tdd-gate`) |
-| `any_type_config.py` | Config for the `Any`-type escape-hatch check |
-| `ask_user_question_shape.py` | Pure AskUserQuestion lean-block shape analyzer: grades question and option-description prose against shared chat-detail markers and sentence/word caps with no hook I/O |
-| `banned_identifiers_constants.py` | The set of banned short identifiers and banned function-name prefixes |
-| `bash_pre_tool_use_dispatcher_constants.py` | Permission outcomes, tool-name sets, and the ordered hosted-hook roster for the Bash PreToolUse dispatcher (also covers PowerShell-shared gates) |
-| `blocking_check_limits.py` | Max issue counts and preview lengths for blocking hooks |
-| `bot_mention_comment_blocker_constants.py` | Patterns for detecting bot @-mentions in PR comments |
-| `claude_md_orphan_file_blocker_constants.py` | Table patterns, file extensions, scan budget, and block-message text for the CLAUDE.md orphan-file blocker |
-| `code_rules_enforcer_constants.py` | File-extension sets, test-path patterns, advisory line thresholds, boolean-name prefixes |
-| `enter_worktree_prefetch_constants.py` | Tool name, origin-remote ref names, and git-command timeouts for the EnterWorktree origin-prefetch hook |
-| `env_var_table_code_drift_constants.py` | Table patterns, env-var-name and code-file recognizers, scan budget, and block-message text for the env-var-table code-drift blocker |
-| `code_rules_path_utils_constants.py` | Path-matching helpers used by the code-rules check modules |
-| `code_verifier_spawn_preflight_gate_constants.py` | Subagent type, merge-tree command flags, timeouts, and deny-message text for the code-verifier spawn pre-flight gate |
-| `command_dispatch_constants.py` | Command-word regex, command-key access pattern, tokenization pattern, and anchors for the unanchored command-dispatch meta-gate |
-| `convergence_branch_constants.py` | Branch and worktree naming patterns for the convergence gate |
-| `convergence_gate_blocker_constants.py` | The full-PR-URL and `--repo`/`-R` flag parse patterns that key the convergence gate to the PR its command names |
-| `conventional_pr_title_gate_constants.py` | Bash tool name, gh executable basenames and pr create/edit subcommand tokens, title/repo flag names, semantic-title-CI workflow markers, the Conventional Commits type list and title pattern, the semantic-pull-request action `types:` input pattern, and block-message text for the conventional-PR-title gate |
-| `dead_argparse_argument_constants.py` | Patterns for detecting unused argparse arguments |
-| `dead_config_field_constants.py` | Patterns for detecting unused `*Config` / `*Selectors` dataclass fields |
-| `dead_dataclass_field_constants.py` | Patterns for detecting unused dataclass fields |
-| `dead_module_constant_constants.py` | Patterns for detecting unexported `UPPER_SNAKE` constants in `*_constants.py` modules |
-| `destructive_command_segment_constants.py` | The list of destructive shell command patterns the blocker matches |
-| `docstring_rule_gate_count_blocker_constants.py` | Target rule basename, spelled-out-number lookup, count-clause and `check_*` validator patterns, and block-message text for the docstring-rule gate-count staleness blocker |
-| `duplicate_function_body_constants.py` | Hashing and comparison config for the duplicate-body check |
-| `duplicate_rmtree_helper_blocker_constants.py` | Sanctioned Windows-safe rmtree helper names, the definition pattern, and the exempt-path set for the duplicate-rmtree-helper blocker |
-| `dynamic_stderr_handler.py` | `DynamicStderrHandler` — a logging handler that resolves `sys.stderr` at emit time (for testability) |
-| `eli11_reply_enforcer_constants.py` | Reply-shape thresholds (bullet cap, lead-line window, per-list-line word cap, overpacked list-line cap), the bullet / numbered-step / table-row / link-target / word patterns, the imperative instruction verbs, and the user-facing notice for the ELI11 reply-shape Stop hook |
-| `fable_spawn_gate_constants.py` | The spawn tool-name set, the fable model alias and the model-id segment split pattern, the `FABLE-SPAWN-AUTHORIZED` authorization token, the advisor-protocol document path, payload field names, the deny message, and the deny `additionalContext` recovery text for the fable-tier spawn gate |
-| `gh_pr_author_swap_constants.py` | Constants for the PR-author swap enforcement hooks |
-| `hardcoded_user_path_constants.py` | Patterns for detecting hardcoded home-directory paths |
-| `harness_scratchpad_constants.py` | Fixed path components (`claude` / `claude-` user-directory name and prefix, `scratchpad` leaf name), the PreToolUse session-id payload key, and the `CLAUDE_CODE_SESSION_ID` environment variable name that the code-rules and TDD gates use to recognize an existing harness session scratchpad directory |
-| `hedging_uncertainty_constants.py` | Sentence-boundary pattern, explicit uncertainty label patterns, hedge-term list separator, and positive corrective block text for OP-07C sentence-scoped hedging |
-| `hook_block_logger.py` | `log_hook_block()` — shared fail-safe logger every blocking hook calls to append a JSON record of each block decision to `~/.claude/logs/hook-blocks.log` |
-| `hook_log_extractor_constants.py` | Neon table name, offset state file path, timeouts, and outcome-type mapping for the hook-log extractor |
-| `hook_prose_detector_consistency_constants.py` | Trigger patterns and corrective messages for the hook-prose consistency checker |
-| `hosted_hook_runner.py` | `run_hook_capturing_output()` — shared runner that runs one dispatcher-hosted hook in-process via runpy and returns its captured stdout and crash flag |
-| `inline_tuple_string_magic_constants.py` | Patterns for detecting magic strings in inline tuple literals |
-| `js_conventions_constants.py` | Banned identifier set, boolean-prefix pattern, and declaration/JSDoc patterns for the JavaScript convention checks |
-| `local_identity.py` | Loader for local machine values: NAS host, ssh user, and ssh port the ssh enforcer guards (`CLAUDE_NAS_*` / `~/.claude/local-identity.json` with placeholder defaults), the PII commit-scan exempt-repo slug set (`CLAUDE_PII_EXEMPT_REPOS` / `pii_exempt_repositories`), and the per-repository allowlisted-values mapping (`pii_allowlisted_values`, keyed by owner/repo slug) read from the git-ignored local file whose path `CLAUDE_LOCAL_IDENTITY_PATH` may point elsewhere; also composes the ssh enforcer's two deny messages that quote the NAS values |
-| `messages.py` | Short user-facing notice strings shown when a Stop hook redirects agent behavior |
-| `multi_edit_reconstruction.py` | `apply_edits()` / `edits_for_tool()` — shared helpers that reconstruct the post-edit content of an Edit or MultiEdit, imported by the blockers that judge post-edit content |
-| `mypy_integration_constants.py` | Path markers (``.git``, ``.py``, ``pyproject.toml``) for mypy project-root resolution |
-| `mypy_validator_cache_constants.py` | Cache paths and tunables for the mypy_validator per-session caches |
-| `nas_ssh_binary_enforcer_constants.py` | Bash tool name, ssh-family basenames, OpenSSH binary path suffixes, and the batch-mode pattern for the NAS ssh binary enforcer; segment helpers come from `shell_command_segments.py` |
-| `open_questions_in_plans_blocker_constants.py` | Patterns for detecting unresolved open questions in plan documents |
-| `orphan_css_class_constants.py` | Scan radius and selector patterns for the orphan-CSS-class check |
-| `package_inventory_stale_blocker_constants.py` | Inventory document names, production code extensions, backtick token pattern, smallest inventory size, exempt names, scan budget, and block-message text for the package-inventory stale-entry blocker |
-| `paired_test_coverage_constants.py` | Test-directory name, stem-test filename affixes, test-file globs, exempt public-function names, scan budget, coverage threshold, and guidance text for the public-function paired-test coverage check |
-| `path_rewriter_constants.py` | Path rewriting patterns for the Everything-search path rewriter |
-| `pii_prevention_constants.py` | Patterns, allowlists, path exemptions, and deny-message text for `pii_prevention_blocker` |
-| `piped_pytest_blocker_constants.py` | Bash tool name, pytest program basenames and the python-interpreter basename pattern, the pipe and segment-reset operator token sets, and the deny message for the piped-pytest blocker; classification that consumes those program and wrapper constants lives in `pytest_invocation.py` |
-| `pytest_invocation.py` | Pure parsed-segment classifier for pytest runs through interpreters, shims, wrappers, and shell string-exec forms; no pipeline deny decision |
-| `plain_language_blocker_constants.py` | The list of heavy words and their everyday replacements |
-| `pr_converge_bugteam_enforcer_constants.py` | State keys and timing config for the bugteam-parallel enforcer |
-| `pr_converge_bugteam_enforcer_state.py` | State-file helpers for the bugteam enforcer |
-| `pr_description_enforcer_constants.py` | PR-description shape rules and command patterns |
-| `pr_description_proof_of_work_constants.py` | Proof-part detection patterns, gh command tokens, and corrective messages for the proof-of-work comment audit |
-| `prose_matcher_precision_constants.py` | Sample floor, keep/narrow precision floors, matcher ids, label decisions, advisory log path, emit caps, and context fingerprint length for OP-07B prose-matcher advisory telemetry |
-| `pre_tool_use_stdin.py` | `read_hook_input_dictionary_from_stdin()` — shared stdin parser for PreToolUse hooks |
-| `precommit_code_rules_gate_constants.py` | Scope argument and exit-code constants for the precommit gate |
-| project_paths_reader.py | Loads ~/.claude/project-paths.json and finds a home-bounded git root for registry checks |
-| `pyproject_config_discovery_constants.py` | Table names (``tool`` key, ``mypy``, ``ruff``) for resolving a validator's config from a pyproject.toml `[tool.<name>]` table |
-| `pytest_testpaths_orphan_blocker_constants.py` | Marker filename, section and key names, test-file pattern, search budget, and block-message text for the pytest unregistered-test-directory blocker |
-| `python_style_checks_constants.py` | Command-line argument count and blank-line count between top-level functions for the style validator |
-| `reviewer_spawn_gate_constants.py` | Bash tool name, the sentinel marker, the Copilot and Bugbot trigger markers, the availability-script relative path and override env-var name, and the deny-message template for the reviewer-spawn gate |
-| `ruff_integration_constants.py` | ``NO_COLOR`` / ``FORCE_COLOR`` environment variable names for plain ruff diagnostics |
-| `send_user_file_open_locally_blocker_constants.py` | Tool name, proactive status, and the block message for the open-locally attach blocker |
-| `sensitive_file_protector_constants.py` | Sensitive filename patterns, the committed-template suffixes that earn an exemption, the write/edit tool names, and the deny decision and message template for `sensitive_file_protector` |
-| `session_edit_stage_gate_constants.py` | Tracker filename prefix/suffix, JSON payload key, edit tool name set, session-id sanitize pattern, lock filename suffix and lock-acquire timing, git diff command, commit flag escapes, and deny-message template shared by the session edit stage gate trio |
-| `session_env_cleanup_constants.py` | Stale-age threshold and directory names for the session-env cleanup hook |
-| `session_start_injector.py` | Pure SessionStart context injector: normalize `source` (`startup`/`resume`/`clear`/`compact`), honor enable/timeout, return structured injection results for starter hooks |
-| `session_start_injector_constants.py` | Known SessionStart source values, injection status codes, enable env var, default timeout for the shared injector |
-| `test_session_start_injector.py` | Behavioral tests for every known source, disabled, timeout, and unknown-source paths |
-| `orchestrator_auto_starter_constants.py` | Opt-in env var, enable values, timeout, and SessionStart directive text for the orchestrator auto-starter |
-| `issue_tracker_session_starter_constants.py` | Opt-in env var, enable values, timeout, and SessionStart directive text for the repository-gated issue-tracker starter |
-| `session_handoff_blocker_constants.py` | Trigger phrases for the session-handoff blocker |
-| `setup_project_paths_constants.py` | Encoding policy, BOM marker, and registry meta-key used across multiple hooks |
-| `shell_command_pipeline.py` | Quote-aware shell pipeline parsing: operator-paired segments after line-continuation join, comment strip, heredoc body drop, and parenthesis-group join |
-| `shell_command_segments.py` | Shell tokenization helpers shared by the segment-walking blockers: segment splitting on control operators, a segment's effective leading program, and a token's basename |
-| `shell_substitution_blocker_constants.py` | Tool name, payload keys, the four substitution-detection patterns, the deny decision shape, and the corrective message for the shell-substitution blocker |
-| `stale_comment_reference_blocker_constants.py` | Identifier pattern, comment stopwords, and denial text for the stale-comment-reference blocker |
-| `state_description_blocker_constants.py` | The set of historical/comparative phrases the state-description blocker rejects, plus the docstring-extraction and mention-span patterns for its Python docstring scan |
-| `stop_dispatcher_constants.py` | Ordered hosted-hook roster and Stop block payload field names for the Stop-chain dispatcher |
-| `stuttering_check_config.py` | Config for the stuttering (repeated-phrase) check |
-| `stuttering_import_binding_constants.py` | Import-binding patterns for the stuttering check |
-| `subprocess_budget_completeness_constants.py` | Required argument names for the subprocess-budget completeness check |
-| `sys_path_insert_constants.py` | Patterns for detecting unguarded `sys.path.insert` calls |
-| `task_list_loop_starter_constants.py` | The one-line task-list instruction and the full session-start directive text for the task-list loop starter hook |
-| `test_layout_constants.py` | The named constants for the test-layout write-time checks |
-| `text_stripping.py` | `strip_code_and_quotes()` — shared helper that removes fenced code blocks, inline code, and blockquotes from prose, imported by the Stop-hook prose blockers |
-| `unscoped_search_blocker_constants.py` | Tool names, `find` and listing program basenames, Get-ChildItem option sets, unscoped-root patterns, the flag/value token stride, and the deny message for the unscoped-search blocker |
-| `unused_module_import_constants.py` | Patterns for detecting unused module-level imports |
-| `volatile_path_in_post_blocker_constants.py` | Volatile path markers, affected `gh` post subcommands, MCP body param names, and the corrective message for the volatile-path post blocker |
-| `windows_rmtree_blocker_constants.py` | The unsafe `shutil.rmtree` pattern and the safe replacement pattern |
-| `workflow_substitution_slot_blocker_constants.py` | Per-iteration token patterns for the workflow-slot blocker |
-| `working_style_prompt_constants.py` | The fixed multi-paragraph working-style prompt text for the SessionStart working_style_prompt hook |
-
-## Conventions
-
-- Most files in this package are pure constants modules — no side effects, no I/O.
-- Hooks import from this package with `from hooks_constants.<module> import <CONSTANT>`.
-- Tests for these modules live beside them as `test_<module>.py`. Run with `python -m pytest hooks_constants/test_<name>.py`.
-- `dynamic_stderr_handler.py`, `pre_tool_use_stdin.py`, `multi_edit_reconstruction.py`, `hosted_hook_runner.py`, and `text_stripping.py` are utility modules (not pure constants) but live here because they hold logic shared across hook surfaces.
-- `local_identity.py` is a loader: it reads the environment or `~/.claude/local-identity.json` (its path overridable via `CLAUDE_LOCAL_IDENTITY_PATH`) to resolve private NAS values (with committed placeholder defaults), the PII commit-scan exempt-repo slug set (`CLAUDE_PII_EXEMPT_REPOS` / `pii_exempt_repositories`), and the per-repository allowlisted-values mapping (`pii_allowlisted_values`).
-
+@AGENTS.md

package/hooks/hooks_constants/bash_pre_tool_use_dispatcher_constants.py

@@ -67,21 +67,10 @@ ALL_BASH_HOSTED_HOOK_ENTRIES: tuple[BashHostedHookEntry, ...] = (
         "blocking/pii_prevention_blocker.py", ALL_BASH_AND_POWERSHELL_TOOL_NAMES
     ),
     BashHostedHookEntry("blocking/conventional_pr_title_gate.py", ALL_BASH_ONLY_TOOL_NAMES),
-    BashHostedHookEntry("blocking/reviewer_spawn_gate.py", ALL_BASH_ONLY_TOOL_NAMES),
     BashHostedHookEntry("blocking/block_main_commit.py", ALL_BASH_ONLY_TOOL_NAMES),
     BashHostedHookEntry("blocking/precommit_code_rules_gate.py", ALL_BASH_ONLY_TOOL_NAMES),
     BashHostedHookEntry("blocking/session_edit_stage_gate.py", ALL_BASH_ONLY_TOOL_NAMES),
-    BashHostedHookEntry("blocking/pr_description_enforcer.py", ALL_BASH_ONLY_TOOL_NAMES),
     BashHostedHookEntry("blocking/test_preflight_check.py", ALL_BASH_ONLY_TOOL_NAMES),
-    BashHostedHookEntry("blocking/convergence_gate_blocker.py", ALL_BASH_ONLY_TOOL_NAMES),
     BashHostedHookEntry("blocking/windows_rmtree_blocker.py", ALL_BASH_ONLY_TOOL_NAMES),
     BashHostedHookEntry("blocking/gh_pr_author_enforcer.py", ALL_BASH_ONLY_TOOL_NAMES),
-    BashHostedHookEntry("blocking/code_review_push_gate.py", ALL_BASH_AND_POWERSHELL_TOOL_NAMES),
-    BashHostedHookEntry(
-        "blocking/code_review_pr_create_gate.py", ALL_BASH_AND_POWERSHELL_TOOL_NAMES
-    ),
-    BashHostedHookEntry(
-        "blocking/code_review_stamp_directory_write_blocker.py",
-        ALL_BASH_AND_POWERSHELL_TOOL_NAMES,
-    ),
 )

package/hooks/hooks_constants/convergence_branch_constants.py

@@ -1,7 +1,6 @@
 """Convergence branch naming policy constants.
 
 Shared by destructive_command_blocker convergence-branch exemptions
-and convergence_gate_blocker pre-flight checks.
 """
 
 ALL_CONVERGENCE_BRANCH_PREFIXES: tuple[str, ...] = ("claude/", "worktree-")

package/hooks/hooks_constants/fable_spawn_gate_constants.py

@@ -11,16 +11,15 @@ into a retry prompt authorizes nothing.
     claude-sonnet-4 -> ['claude', 'sonnet', '4']  ok:   another tier
 
 Every literal the gate body reads lives here. ``AGENT_TOOL_NAME`` and
-``TASK_TOOL_NAME`` come from the sibling spawn-gate constants, so the spawn
-gates share one source for both tool names.
+``TASK_TOOL_NAME`` name the spawn tools the gate evaluates.
 """
 
 from __future__ import annotations
 
 import re
 
-from hooks_constants.code_verifier_spawn_preflight_gate_constants import TASK_TOOL_NAME
-from hooks_constants.pr_converge_bugteam_enforcer_constants import AGENT_TOOL_NAME
+AGENT_TOOL_NAME: str = "Agent"
+TASK_TOOL_NAME: str = "Task"
 
 ALL_SPAWN_TOOL_NAMES: frozenset[str] = frozenset({AGENT_TOOL_NAME, TASK_TOOL_NAME})
 

package/hooks/hooks_constants/hook_log_extractor_constants.py

@@ -123,7 +123,7 @@ SUMMARY_COLUMN_HEADINGS: tuple[str, str, str, str] = (
     "top_blocked_command_preview",
 )
 
-SUMMARY_NO_NEW_BLOCKS_MESSAGE: str = "No new blocks since last run."
+SUMMARY_ZERO_BLOCKED_EVENTS_MESSAGE: str = "Blocked events recorded in the last 24 hours: 0."
 
 QUERY_NO_ROWS_RETURNED_MESSAGE: str = "No rows returned."
 
@@ -231,4 +231,3 @@ WINDOWS_CREATE_NEW_PROCESS_GROUP_FLAG: int = 0x00000200
 
 LOCK_MAXIMUM_RETRY_COUNT: int = 30
 LOCK_RETRY_SLEEP_SECONDS: float = 0.1
-

package/hooks/hooks_constants/pre_tool_use_dispatcher_constants.py

@@ -86,10 +86,6 @@ ALL_HOSTED_HOOK_ENTRIES: tuple[HostedHookEntry, ...] = (
         script_relative_path="blocking/sensitive_file_protector.py",
         applicable_tool_names=ALL_WRITE_AND_EDIT_TOOL_NAMES,
     ),
-    HostedHookEntry(
-        script_relative_path="blocking/code_review_stamp_directory_write_blocker.py",
-        applicable_tool_names=ALL_WRITE_EDIT_MULTI_EDIT_TOOL_NAMES,
-    ),
     HostedHookEntry(
         script_relative_path="blocking/pii_prevention_blocker.py",
         applicable_tool_names=ALL_WRITE_EDIT_MULTI_EDIT_TOOL_NAMES,
@@ -131,10 +127,6 @@ ALL_HOSTED_HOOK_ENTRIES: tuple[HostedHookEntry, ...] = (
         script_relative_path="blocking/hook_prose_detector_consistency.py",
         applicable_tool_names=ALL_WRITE_AND_EDIT_TOOL_NAMES,
     ),
-    HostedHookEntry(
-        script_relative_path="blocking/verified_commit_message_accuracy_blocker.py",
-        applicable_tool_names=ALL_WRITE_AND_EDIT_TOOL_NAMES,
-    ),
     HostedHookEntry(
         script_relative_path="blocking/workflow_substitution_slot_blocker.py",
         applicable_tool_names=ALL_WRITE_EDIT_MULTI_EDIT_TOOL_NAMES,

package/hooks/hooks_constants/send_user_file_open_locally_blocker_constants.py

@@ -7,12 +7,7 @@ PROACTIVE_STATUS: str = "proactive"
 CORRECTIVE_MESSAGE: str = (
     "BLOCKED [open-locally]: SendUserFile attaches a file to the session, which "
     "does not let the user see it while they are at the terminal. Open the file on "
-    "screen in its own viewer:\n"
-    "  Start-Process pwsh -WindowStyle Hidden -ArgumentList "
-    "'-NoProfile','-File',\"$HOME\\.claude\\scripts\\Show-Asset.ps1\","
-    "'<path 1>','<path 2>'\n"
-    "Show-Asset.ps1 sizes each image window to the image and opens every other file "
-    "type in its default app. Pass every path the user named.\n"
+    "screen with its native Windows app. Open every path the user named.\n"
     "The one allowed attach is a phone push: when the user has stepped away and you "
     'want the file to reach their phone, call SendUserFile with status "proactive".'
 )

package/hooks/hooks_constants/stop_dispatcher_constants.py

@@ -23,5 +23,4 @@ ALL_STOP_HOSTED_HOOK_PATHS: tuple[str, ...] = (
     "blocking/intent_only_ending_blocker.py",
     "blocking/session_handoff_blocker.py",
     "blocking/eli11_reply_enforcer.py",
-    "diagnostic/hook_log_stop_wrapper.py",
 )

package/hooks/hooks_constants/task_list_loop_starter_constants.py

@@ -11,8 +11,9 @@ TASK_LIST_MAINTENANCE_INSTRUCTION = "build or update the task list using the tas
 
 TASK_LIST_LOOP_DIRECTIVE = (
     "TASK LIST MAINTENANCE: Keep this session's task list current on a 10-minute "
-    "cadence. If a task-list maintenance loop is not already running this session, "
-    "start one now with Claude Code's built-in /loop skill on a 10-minute interval: "
+    "cadence. Reuse the active task-list maintenance loop when available. Start "
+    "Claude Code's built-in /loop skill with a 10-minute interval when this session "
+    "needs its first loop: "
     f'/loop 10m "{TASK_LIST_MAINTENANCE_INSTRUCTION}". Run that instruction once '
-    "immediately, then let the loop refresh the list every 10 minutes."
+    "immediately when starting the loop, then let the loop refresh the list every 10 minutes."
 )

package/hooks/hooks_constants/test_bash_pre_tool_use_dispatcher_constants.py

@@ -32,26 +32,17 @@ _EXPECTED_BASH_ORDER = (
     "blocking/volatile_path_in_post_blocker.py",
     "blocking/pii_prevention_blocker.py",
     "blocking/conventional_pr_title_gate.py",
-    "blocking/reviewer_spawn_gate.py",
     "blocking/block_main_commit.py",
     "blocking/precommit_code_rules_gate.py",
     "blocking/session_edit_stage_gate.py",
-    "blocking/pr_description_enforcer.py",
     "blocking/test_preflight_check.py",
-    "blocking/convergence_gate_blocker.py",
     "blocking/windows_rmtree_blocker.py",
     "blocking/gh_pr_author_enforcer.py",
-    "blocking/code_review_push_gate.py",
-    "blocking/code_review_pr_create_gate.py",
-    "blocking/code_review_stamp_directory_write_blocker.py",
 )
 
 _POWERSHELL_APPLICABLE = (
     "blocking/unscoped_search_blocker.py",
     "blocking/pii_prevention_blocker.py",
-    "blocking/code_review_push_gate.py",
-    "blocking/code_review_pr_create_gate.py",
-    "blocking/code_review_stamp_directory_write_blocker.py",
 )
 
 
@@ -70,7 +61,7 @@ def test_every_hook_applies_to_the_bash_tool() -> None:
 
 
 def test_powershell_applicable_hooks_include_the_shared_gates() -> None:
-    """PowerShell runs the unscoped-search and PII gates plus the review family."""
+    """PowerShell runs the shared unscoped-search and PII gates."""
     powershell_hooks = tuple(
         each_entry.script_relative_path
         for each_entry in ALL_BASH_HOSTED_HOOK_ENTRIES
@@ -86,22 +77,6 @@ def test_powershell_hooks_carry_the_shared_tool_set() -> None:
             assert each_entry.applicable_tool_names == ALL_BASH_AND_POWERSHELL_TOOL_NAMES
 
 
-def test_code_review_gate_family_registered_for_bash_and_powershell() -> None:
-    """The three code-review shell gates run on both Bash and PowerShell."""
-    code_review_gate_paths = (
-        "blocking/code_review_push_gate.py",
-        "blocking/code_review_pr_create_gate.py",
-        "blocking/code_review_stamp_directory_write_blocker.py",
-    )
-    entry_by_path = {
-        each_entry.script_relative_path: each_entry
-        for each_entry in ALL_BASH_HOSTED_HOOK_ENTRIES
-    }
-    for each_gate_path in code_review_gate_paths:
-        assert each_gate_path in entry_by_path
-        assert entry_by_path[each_gate_path].applicable_tool_names == (
-            ALL_BASH_AND_POWERSHELL_TOOL_NAMES
-        )
 
 
 def test_every_roster_path_points_at_an_existing_hook() -> None:

package/hooks/hooks_constants/test_pre_tool_use_dispatcher_constants.py

@@ -63,11 +63,6 @@ def test_duplicate_rmtree_helper_blocker_runs_via_runpy() -> None:
     assert entry.native_module_name is None
 
 
-def test_code_review_stamp_write_blocker_registered_for_all_file_tools() -> None:
-    entry = _entry_for("blocking/code_review_stamp_directory_write_blocker.py")
-    assert entry is not None
-    assert {"Write", "Edit", "MultiEdit"} <= entry.applicable_tool_names
-
 
 def test_windows_rmtree_blocker_still_registered() -> None:
     entry = _entry_for("blocking/windows_rmtree_blocker.py")

package/hooks/hooks_constants/test_stop_dispatcher_constants.py

@@ -17,7 +17,6 @@ _EXPECTED_STOP_ORDER = (
     "blocking/intent_only_ending_blocker.py",
     "blocking/session_handoff_blocker.py",
     "blocking/eli11_reply_enforcer.py",
-    "diagnostic/hook_log_stop_wrapper.py",
 )
 
 

package/hooks/hooks_constants/working_style_prompt_constants.py

@@ -7,24 +7,33 @@ __all__ = [
 ]
 
 WORKING_STYLE_PROMPT = (
-    "document everything in a location you can find later, rather than relying on "
-    "memory. scratch txt file you'll keep running as you go; ledger, if you will."
-    "Always, without exception, write in plain english. Keep responses focused, "
-    "brief, and concise. Keep disclaimers and caveats short, and spend most of the "
-    "response on the main answer. When asked to explain something, give a high-level "
-    "summary unless an in-depth explanation is specifically requested."
-    "Before your first tool call, say in one sentence what you're about to do. While "
-    "working, give a brief update only when you find something important or change "
-    "direction. When you finish, lead with the outcome: your first sentence should "
-    "answer \"what happened\" or \"what did you find,\" with supporting detail after "
-    "it for readers who want it."
-    "Match the length of written documents to what the task needs: cover the "
-    "substance, but do not pad with filler sections, redundant summaries, or "
-    "boilerplate."
-    "Deliver what was asked, at the scope intended. Make routine judgment calls "
-    "yourself, and check in only when different readings of the request would lead "
-    "to materially different work. If the request seems mistaken or a better "
-    "approach exists, say so in a sentence and continue with the task as asked "
-    "rather than quietly narrowing, widening, or transforming it. Finish the whole "
-    "task, and stop short of actions that are clearly beyond what was asked."
+    "Document each task in a location that remains easy to find later. Keep a "
+    "running scratch text ledger as you work. Write in plain English. Keep responses "
+    "focused, brief, and concise. Keep disclaimers and caveats short while giving "
+    "the main answer most of the response. Give a high-level explanation by default "
+    "and provide depth when the request calls for it. "
+    "Before your first tool call, state your next action in one sentence. While "
+    "working, give brief updates when you find important information or change "
+    "direction. Finish with the outcome in the first sentence, then provide "
+    "supporting detail for readers who want it. "
+    "Match written-document length to the task. Cover the substance and keep every "
+    "section, summary, and phrase useful. "
+    "Deliver the requested work at its intended scope. Make routine judgment calls "
+    "yourself. Ask for direction when different interpretations would produce "
+    "materially different work. When a request seems mistaken or a better approach "
+    "exists, state the concern briefly and continue with the requested task. Finish "
+    "the complete task and keep actions within the requested scope. "
+    "When a request has multiple reasonable interpretations, state your understanding "
+    "and the assumptions that shape the work. Ask one focused clarification question "
+    "when the ambiguity changes the outcome, scope, audience, format, or risk. Use a "
+    "clearly stated low-risk assumption when the intended result remains stable. "
+    "Pause for the user's choice before making a high-impact decision. "
+    "Use positive prose throughout every generated text surface, including "
+    "documentation, comments, user-facing messages, sub-agent prompts, labels, "
+    "plans, and instructions. Write each point as one direct affirmative statement "
+    "that presents only the intended action, fact, reason, or outcome. Use current, "
+    "immediately relevant context and omit historical clutter. Use plain language, "
+    "full terms, and simple descriptive names that communicate their meaning "
+    "immediately. Keep all text concise, clear, direct, and focused entirely on "
+    "useful information."
 )

package/hooks/lifecycle/AGENTS.md

@@ -0,0 +1,18 @@
+# hooks/lifecycle
+
+Hooks that run at session or config-change boundaries rather than on individual tool calls.
+
+## Key files
+
+| File | Event | What it does |
+|---|---|---|
+| `config_change_guard.py` | PostToolUse (Write/Edit on `settings.json`) | Counts hooks in the edited `settings.json` and logs any change to `~/.claude/cache/config-change-audit.log`; alerts when the hook count drops below the last known value |
+| `enter_worktree_origin_prefetch.py` | PreToolUse (EnterWorktree) | Fetches origin's default branch before a worktree creation call, keeping the `refs/remotes/origin/<default-branch>` ref that `fresh` mode reads current; never blocks on fetch failure |
+| `session_end_cleanup.py` | SessionEnd | Purges stale cache entries from `~/.claude/cache/` (entries older than the configured threshold) and old backup files |
+| `test_config_change_guard.py` | — | Tests for `config_change_guard.py` |
+| `test_enter_worktree_origin_prefetch.py` | — | Tests for `enter_worktree_origin_prefetch.py` |
+
+## Conventions
+
+- Constants for these hooks (stale-age threshold, cache directory, known-hook-count file, EnterWorktree prefetch tuning) live in `hooks_constants/session_env_cleanup_constants.py` and `hooks_constants/enter_worktree_prefetch_constants.py`.
+- Tests run with `python -m pytest lifecycle/test_<name>.py`.

package/hooks/lifecycle/CLAUDE.md

@@ -1,20 +1 @@
-# hooks/lifecycle
-
-Hooks that run at session or config-change boundaries rather than on individual tool calls.
-
-## Key files
-
-| File | Event | What it does |
-|---|---|---|
-| `config_change_guard.py` | PostToolUse (Write/Edit on `settings.json`) | Counts hooks in the edited `settings.json` and logs any change to `~/.claude/cache/config-change-audit.log`; alerts when the hook count drops below the last known value |
-| `enter_worktree_origin_prefetch.py` | PreToolUse (EnterWorktree) | Fetches origin's default branch before a worktree creation call, keeping the `refs/remotes/origin/<default-branch>` ref that `fresh` mode reads current; never blocks on fetch failure |
-| `pr_converge_bugteam_skill_tracker.py` | PreToolUse (Skill) | Tracks which bugteam skill runs have completed within a pr-converge loop, so the enforcer can verify parallel execution |
-| `session_end_cleanup.py` | SessionEnd | Purges stale cache entries from `~/.claude/cache/` (entries older than the configured threshold) and old backup files |
-| `test_config_change_guard.py` | — | Tests for `config_change_guard.py` |
-| `test_enter_worktree_origin_prefetch.py` | — | Tests for `enter_worktree_origin_prefetch.py` |
-| `test_pr_converge_bugteam_skill_tracker.py` | — | Tests for `pr_converge_bugteam_skill_tracker.py` |
-
-## Conventions
-
-- Constants for these hooks (stale-age threshold, cache directory, known-hook-count file, EnterWorktree prefetch tuning) live in `hooks_constants/session_env_cleanup_constants.py`, `hooks_constants/pr_converge_bugteam_enforcer_constants.py`, and `hooks_constants/enter_worktree_prefetch_constants.py`.
-- Tests run with `python -m pytest lifecycle/test_<name>.py`.
+@AGENTS.md

package/hooks/observability/AGENTS.md

@@ -0,0 +1,20 @@
+# hooks/observability
+
+PostToolUse hooks that record agent behavior for later review. These hooks do not block tool calls; they log or annotate them.
+
+## Key files
+
+| File | Event | What it does |
+|---|---|---|
+| `instructions_loaded_logger.py` | PostToolUse (file load events) | Appends a JSONL record to `~/.claude/logs/instructions_loaded.jsonl` each time Claude Code loads a context file (CLAUDE.md, rules, skills), capturing the file path, load reason, memory type, and session ID |
+| `session_file_edit_tracker.py` | PostToolUse (Write/Edit/MultiEdit) | Records the resolved absolute path of each file edited this session into a per-session JSON file under the system temp directory, for the session edit stage gate to read at commit time |
+| `prose_matcher_advisory.py` | Advisory emit from prose matchers | Privacy-safe JSONL candidates (`matcher_id`, surface, context fingerprint, label) plus keep/narrow/drop/advisory classification against the sample floor |
+| `test_instructions_loaded_logger.py` | — | Tests for `instructions_loaded_logger.py` |
+| `test_session_file_edit_tracker.py` | — | Tests for `session_file_edit_tracker.py` |
+| `test_prose_matcher_advisory.py` | — | Tests for advisory emit and precision classification |
+
+## Conventions
+
+- The log file is append-only; the hook creates the parent directory if needed.
+- Errors during logging are caught and written as error records rather than propagated — the hook never blocks the tool call.
+- Tests run with `python -m pytest observability/test_instructions_loaded_logger.py`.

package/hooks/observability/CLAUDE.md

@@ -1,20 +1 @@
-# hooks/observability
-
-PostToolUse hooks that record agent behavior for later review. These hooks do not block tool calls; they log or annotate them.
-
-## Key files
-
-| File | Event | What it does |
-|---|---|---|
-| `instructions_loaded_logger.py` | PostToolUse (file load events) | Appends a JSONL record to `~/.claude/logs/instructions_loaded.jsonl` each time Claude Code loads a context file (CLAUDE.md, rules, skills), capturing the file path, load reason, memory type, and session ID |
-| `session_file_edit_tracker.py` | PostToolUse (Write/Edit/MultiEdit) | Records the resolved absolute path of each file edited this session into a per-session JSON file under the system temp directory, for the session edit stage gate to read at commit time |
-| `prose_matcher_advisory.py` | Advisory emit from prose matchers | Privacy-safe JSONL candidates (`matcher_id`, surface, context fingerprint, label) plus keep/narrow/drop/advisory classification against the sample floor |
-| `test_instructions_loaded_logger.py` | — | Tests for `instructions_loaded_logger.py` |
-| `test_session_file_edit_tracker.py` | — | Tests for `session_file_edit_tracker.py` |
-| `test_prose_matcher_advisory.py` | — | Tests for advisory emit and precision classification |
-
-## Conventions
-
-- The log file is append-only; the hook creates the parent directory if needed.
-- Errors during logging are caught and written as error records rather than propagated — the hook never blocks the tool call.
-- Tests run with `python -m pytest observability/test_instructions_loaded_logger.py`.
+@AGENTS.md

package/hooks/session/AGENTS.md

@@ -0,0 +1,33 @@
+# hooks/session
+
+SessionStart and SessionEnd hooks for per-session setup and cleanup: removing stale session and plugin-data directories at startup, detecting unregistered repositories, starting the session's task-list maintenance loop, injecting working-style guidance, and clearing PR-author swap state at shutdown.
+
+## Key files
+
+| File | Event | What it does |
+|---|---|---|
+| `session_env_cleanup.py` | SessionStart | Removes the current session's pre-existing `~/.claude/session-env/<session_id>/` directory and prunes sibling entries older than the stale-age threshold. Prevents `EEXIST` errors from non-recursive `mkdir` calls in the Bash tool on Windows. |
+| `gh_pr_author_session_cleanup.py` | SessionEnd | Clears any PR-author swap state left over from the current session |
+| `session_edit_tracker_cleanup.py` | SessionStart, SessionEnd | Deletes the tracker file for the running Claude Code conversation from the system temp directory — at start for a clean slate and at end for a clean exit. A tracker is read only by the conversation that wrote it, so a live idle tracker is kept while a peer cleans up |
+| `plugin_data_dir_cleanup.py` | SessionStart | Removes empty plugin data directories at startup to prevent `EEXIST` when Claude Code recreates them |
+| `untracked_repo_detector.py` | SessionStart | Detects when the session cwd is inside a git repository that is not registered in `~/.claude/project-paths.json` and logs a warning |
+| `task_list_loop_starter.py` | SessionStart | Emits an `additionalContext` directive telling Claude to keep the task list current on a 10-minute cadence, starting the `/loop` skill when one is not already running. Writes nothing and runs no tools itself. |
+| `orchestrator_auto_starter.py` | SessionStart | Opt-in (`CLAUDE_ORCHESTRATOR_AUTO_STARTER_ENABLED`) consumer of the shared SessionStart injector; emits orchestrator skill context when enabled. Manual `/orchestrator` unchanged. |
+| `issue_tracker_session_starter.py` | SessionStart | Opt-in (`CLAUDE_ISSUE_TRACKER_SESSION_STARTER_ENABLED`) and repository-gated (git root in `~/.claude/project-paths.json`) issue-tracker skill context. |
+| `_path_setup.py` | — | Inserts the hooks directory on `sys.path` so SessionStart scripts import `hooks_constants` with top-level imports. |
+| `test_orchestrator_auto_starter.py` | — | Tests for `orchestrator_auto_starter.py` |
+| `test_issue_tracker_session_starter.py` | — | Tests for `issue_tracker_session_starter.py` |
+| `working_style_prompt.py` | SessionStart | Emits an `additionalContext` block with the fixed working-style prompt (running ledger, plain English, outcome-first finish, scope discipline). Writes nothing and runs no tools itself. |
+| `test_gh_pr_author_session_cleanup.py` | — | Tests for `gh_pr_author_session_cleanup.py` |
+| `test_session_edit_tracker_cleanup.py` | — | Tests for `session_edit_tracker_cleanup.py` |
+| `test_session_env_cleanup.py` | — | Tests for `session_env_cleanup.py` |
+| `test_untracked_repo_detector.py` | — | Tests for `untracked_repo_detector.py` |
+| `test_task_list_loop_starter.py` | — | Tests for `task_list_loop_starter.py` |
+| `test__path_setup.py` | — | Tests for `_path_setup.py` |
+| `test_working_style_prompt.py` | — | Tests for `working_style_prompt.py` |
+
+## Conventions
+
+- `session_env_cleanup.py` is Windows-specific in effect but safe to run on all platforms; it exits 0 when the target directory does not exist.
+- Constants (stale-age threshold, directory names) live in `hooks_constants/session_env_cleanup_constants.py`.
+- Tests run with `python -m pytest session/test_<name>.py`.

package/hooks/session/CLAUDE.md

@@ -1,33 +1 @@
-# hooks/session
-
-SessionStart and SessionEnd hooks for per-session setup and cleanup: removing stale session and plugin-data directories at startup, detecting unregistered repositories, starting the session's task-list maintenance loop, injecting working-style guidance, and clearing PR-author swap state at shutdown.
-
-## Key files
-
-| File | Event | What it does |
-|---|---|---|
-| `session_env_cleanup.py` | SessionStart | Removes the current session's pre-existing `~/.claude/session-env/<session_id>/` directory and prunes sibling entries older than the stale-age threshold. Prevents `EEXIST` errors from non-recursive `mkdir` calls in the Bash tool on Windows. |
-| `gh_pr_author_session_cleanup.py` | SessionEnd | Clears any PR-author swap state left over from the current session |
-| `session_edit_tracker_cleanup.py` | SessionStart, SessionEnd | Deletes the tracker file for the running Claude Code conversation from the system temp directory — at start for a clean slate and at end for a clean exit. A tracker is read only by the conversation that wrote it, so a live idle tracker is kept while a peer cleans up |
-| `plugin_data_dir_cleanup.py` | SessionStart | Removes empty plugin data directories at startup to prevent `EEXIST` when Claude Code recreates them |
-| `untracked_repo_detector.py` | SessionStart | Detects when the session cwd is inside a git repository that is not registered in `~/.claude/project-paths.json` and logs a warning |
-| `task_list_loop_starter.py` | SessionStart | Emits an `additionalContext` directive telling Claude to keep the task list current on a 10-minute cadence, starting the `/loop` skill when one is not already running. Writes nothing and runs no tools itself. |
-| `orchestrator_auto_starter.py` | SessionStart | Opt-in (`CLAUDE_ORCHESTRATOR_AUTO_STARTER_ENABLED`) consumer of the shared SessionStart injector; emits orchestrator skill context when enabled. Manual `/orchestrator` unchanged. |
-| `issue_tracker_session_starter.py` | SessionStart | Opt-in (`CLAUDE_ISSUE_TRACKER_SESSION_STARTER_ENABLED`) and repository-gated (git root in `~/.claude/project-paths.json`) issue-tracker skill context. |
-| `_path_setup.py` | — | Inserts the hooks directory on `sys.path` so SessionStart scripts import `hooks_constants` with top-level imports. |
-| `test_orchestrator_auto_starter.py` | — | Tests for `orchestrator_auto_starter.py` |
-| `test_issue_tracker_session_starter.py` | — | Tests for `issue_tracker_session_starter.py` |
-| `working_style_prompt.py` | SessionStart | Emits an `additionalContext` block with the fixed working-style prompt (running ledger, plain English, outcome-first finish, scope discipline). Writes nothing and runs no tools itself. |
-| `test_gh_pr_author_session_cleanup.py` | — | Tests for `gh_pr_author_session_cleanup.py` |
-| `test_session_edit_tracker_cleanup.py` | — | Tests for `session_edit_tracker_cleanup.py` |
-| `test_session_env_cleanup.py` | — | Tests for `session_env_cleanup.py` |
-| `test_untracked_repo_detector.py` | — | Tests for `untracked_repo_detector.py` |
-| `test_task_list_loop_starter.py` | — | Tests for `task_list_loop_starter.py` |
-| `test__path_setup.py` | — | Tests for `_path_setup.py` |
-| `test_working_style_prompt.py` | — | Tests for `working_style_prompt.py` |
-
-## Conventions
-
-- `session_env_cleanup.py` is Windows-specific in effect but safe to run on all platforms; it exits 0 when the target directory does not exist.
-- Constants (stale-age threshold, directory names) live in `hooks_constants/session_env_cleanup_constants.py`.
-- Tests run with `python -m pytest session/test_<name>.py`.
+@AGENTS.md

package/hooks/session/gh_pr_author_session_cleanup.py

@@ -34,6 +34,7 @@ if hooks_parent_directory not in sys.path:
     sys.path.insert(0, hooks_parent_directory)
 
 from _gh_pr_author_swap_utils import (  # noqa: E402
+    _build_restore_failure_message,
     _delete_state_file,
     _lstat_indicates_attacker_planted,
     _read_original_account,
@@ -142,8 +143,12 @@ def _restore_stale_state_file(state_file: Path) -> None:
         _delete_state_file(state_file)
     else:
         _write_line(
-            f"[gh-pr-author-cleanup] failed to restore active gh account to {original_account!r} from "
-            f"stale state file {state_file}; left in place for next session",
+            _build_restore_failure_message(
+                "gh-pr-author-cleanup",
+                original_account,
+                state_file,
+                "stale state file",
+            ),
             sys.stderr,
         )
 

package/hooks/session/task_list_loop_starter.py

@@ -2,9 +2,9 @@
 """SessionStart hook — start a task-list maintenance loop for the session.
 
 At session start this hook emits an ``additionalContext`` directive asking Claude
-to keep the session's task list current on a 10-minute cadence, starting the
-``/loop`` skill when one is not already running. The hook writes nothing and runs
-no tools itself. Claude reads the directive and invokes the ``/loop`` skill.
+to keep the session's task list current on a 10-minute cadence. Claude reuses the
+active ``/loop`` skill when available and starts the skill for the session's first
+task-list maintenance loop.
 """
 
 from __future__ import annotations

package/hooks/session/test_gh_pr_author_session_cleanup.py

@@ -343,7 +343,8 @@ def test_restore_stale_state_file_logs_when_switch_fails(
 
     captured_streams = capsys.readouterr()
     assert state_file.exists()
-    assert "[gh-pr-author-cleanup] failed to restore" in captured_streams.err
+    assert "[gh-pr-author-cleanup] Restore the gh account" in captured_streams.err
+    assert "gh account switch requires attention" in captured_streams.err
     assert "'jl-cmd'" in captured_streams.err
     assert str(state_file) in captured_streams.err