claude-dev-env 2.15.1 → 2.17.0

package/hooks/blocking/eli11_reply_enforcer.py

@@ -1,436 +0,0 @@
-#!/usr/bin/env python3
-"""
-Stop hook that blocks egregious reply-shape failures (action-first, bullets).
-
-The `eli11-replies` and `opus5-communication-contract` rules ask every chat
-reply to lead with the action when the user must act, keep findings to a few
-bullets, and stay scannable. This hook reads the final assistant message,
-strips code fences, inline code, blockquotes, table rows, and link targets,
-then judges shape only: multi-line instructions need numbered steps first,
-bullet count stays bounded, and list lines stay one-idea short.
-"""
-
-import json
-import sys
-from pathlib import Path
-
-_hooks_dir = str(Path(__file__).resolve().parent.parent)
-if _hooks_dir not in sys.path:
-    sys.path.insert(0, _hooks_dir)
-
-from hooks_constants.eli11_reply_enforcer_constants import (  # noqa: E402
-    ACTION_FIRST_LEAD_LINE_COUNT,
-    ALL_IMPERATIVE_INSTRUCTION_VERBS,
-    ALL_IMPERATIVE_OBJECT_LEAD_WORDS,
-    ALPHABETIC_WORD_PATTERN,
-    BULLET_LINE_PATTERN,
-    COUNTABLE_WORD_PATTERN,
-    IMPERATIVE_OBJECT_TOKEN_PATTERN,
-    LINK_TARGET_PATTERN,
-    LIST_MARKER_PREFIX_PATTERN,
-    MARKDOWN_LEAD_MARKER_PATTERN,
-    MAXIMUM_BULLET_LINE_COUNT,
-    MAXIMUM_OVERPACKED_LIST_LINE_COUNT,
-    MAXIMUM_WORDS_PER_LIST_LINE,
-    NUMBERED_STEP_PATTERN,
-    TABLE_ROW_PATTERN,
-    TARGET_BULLET_LINE_COUNT,
-    USER_FACING_ELI11_NOTICE,
-    VIOLATION_SEPARATOR,
-)
-from hooks_constants.hook_block_logger import log_hook_block  # noqa: E402
-from hooks_constants.text_stripping import strip_code_and_quotes  # noqa: E402
-
-
-def extract_reply_prose(assistant_message: str) -> str:
-    """Return the reader-visible prose of a reply.
-
-    ::
-
-        in:  "See https://example.com/a now"
-        out: "See  now"
-
-    Fenced code, inline code, and blockquotes come off through the shared
-    stripper; table rows and link targets come off here.
-
-    Args:
-        assistant_message: The raw final assistant message.
-
-    Returns:
-        The message with code, quotes, table rows, and link targets removed.
-    """
-    prose_text = strip_code_and_quotes(assistant_message)
-    prose_text = TABLE_ROW_PATTERN.sub("", prose_text)
-    return LINK_TARGET_PATTERN.sub("", prose_text)
-
-
-def collect_non_empty_lines(prose_text: str) -> list[str]:
-    """Return every line of the stripped prose that carries visible text.
-
-    Args:
-        prose_text: Prose already cleaned of code, quotes, tables, and links.
-
-    Returns:
-        The lines with content, in order, blank lines dropped.
-    """
-    return [
-        each_line for each_line in prose_text.splitlines() if each_line.strip()
-    ]
-
-
-def leading_word_of(prose_line: str) -> str:
-    """Return the first word of a line once list and bold markers come off.
-
-    ::
-
-        in:  "- **Click** Save"
-        out: "click"
-
-    Args:
-        prose_line: One non-empty line of reply prose.
-
-    Returns:
-        The lowercased first alphabetic word, or an empty string when the line
-        carries no letters.
-    """
-    unmarked_line = LIST_MARKER_PREFIX_PATTERN.sub("", prose_line, count=1)
-    all_leading_words = ALPHABETIC_WORD_PATTERN.findall(unmarked_line)
-    if not all_leading_words:
-        return ""
-    return all_leading_words[0].lower()
-
-
-def object_word_after_leading_verb(prose_line: str) -> str:
-    """Return the word a line puts right after its opening verb.
-
-    ::
-
-        in:  "Run the migration script."
-        out: "the"
-
-    Args:
-        prose_line: One non-empty line of reply prose.
-
-    Returns:
-        The lowercased token following the first alphabetic word, or an empty
-        string when the line carries no such token.
-    """
-    unmarked_line = LIST_MARKER_PREFIX_PATTERN.sub("", prose_line, count=1)
-    leading_word_match = ALPHABETIC_WORD_PATTERN.search(unmarked_line)
-    if leading_word_match is None:
-        return ""
-    all_trailing_tokens = COUNTABLE_WORD_PATTERN.findall(
-        unmarked_line[leading_word_match.end():]
-    )
-    if not all_trailing_tokens:
-        return ""
-    return all_trailing_tokens[0].lower()
-
-
-def names_imperative_object(object_word: str) -> bool:
-    """Return True when a word reads as the thing an imperative acts on.
-
-    ::
-
-        ok:   "the" (Run the script), "3" (Do 3 things), "scripts/run.py"
-        flag: "questions" (Open questions remain), "is" (Merge is complete)
-
-    An imperative names its object through a determiner, a count, a path, or a
-    filename; a narrative sentence puts a bare subject noun there instead.
-
-    Args:
-        object_word: The lowercased word following a line's opening verb.
-
-    Returns:
-        True when the word marks the opening verb as a real imperative.
-    """
-    if not object_word:
-        return False
-    if object_word in ALL_IMPERATIVE_OBJECT_LEAD_WORDS:
-        return True
-    return IMPERATIVE_OBJECT_TOKEN_PATTERN.search(object_word) is not None
-
-
-def is_imperative_instruction_line(prose_line: str) -> bool:
-    """Return True when a line tells the user to act.
-
-    ::
-
-        ok:   "Open questions remain about the stripper." -> False
-        flag: "Open the pull request."                    -> True
-
-    Args:
-        prose_line: One non-empty line of reply prose.
-
-    Returns:
-        True when the line opens with a tracked instruction verb naming an
-        object an imperative acts on.
-    """
-    if leading_word_of(prose_line) not in ALL_IMPERATIVE_INSTRUCTION_VERBS:
-        return False
-    return names_imperative_object(object_word_after_leading_verb(prose_line))
-
-
-def has_imperative_instruction_line(all_prose_lines: list[str]) -> bool:
-    """Return True when any line opens with a verb telling the user to act.
-
-    Args:
-        all_prose_lines: The non-empty lines of reply prose.
-
-    Returns:
-        True when a line reads as an imperative instruction.
-    """
-    return any(
-        is_imperative_instruction_line(each_line) for each_line in all_prose_lines
-    )
-
-
-def has_leading_numbered_step(all_prose_lines: list[str]) -> bool:
-    """Return True when a numbered step appears among the lead lines.
-
-    Args:
-        all_prose_lines: The non-empty lines of reply prose.
-
-    Returns:
-        True when one of the first lead lines opens a numbered list.
-    """
-    all_lead_lines = all_prose_lines[:ACTION_FIRST_LEAD_LINE_COUNT]
-    return any(
-        NUMBERED_STEP_PATTERN.match(each_line) for each_line in all_lead_lines
-    )
-
-
-def count_bullet_lines(all_prose_lines: list[str]) -> int:
-    """Return how many lines of the reply are bullets.
-
-    Args:
-        all_prose_lines: The non-empty lines of reply prose.
-
-    Returns:
-        The count of lines opening with a bullet marker.
-    """
-    return sum(
-        1 for each_line in all_prose_lines if BULLET_LINE_PATTERN.match(each_line)
-    )
-
-
-def count_words_in_line(prose_line: str) -> int:
-    """Return the countable words a line carries once its list marker comes off.
-
-    ::
-
-        in:  "- **Save** the draft"
-        out: 3
-
-    Args:
-        prose_line: One non-empty line of reply prose.
-
-    Returns:
-        The count of tokens carrying a letter or digit, marker text excluded.
-    """
-    unmarked_line = LIST_MARKER_PREFIX_PATTERN.sub("", prose_line, count=1)
-    return len(COUNTABLE_WORD_PATTERN.findall(unmarked_line))
-
-
-def strip_markdown_lead_markers(prose_line: str) -> str:
-    """Return a line with its opening blockquote, heading, and bold markers off.
-
-    ::
-
-        in:  "> **Bold lead:** the report follows"
-        out: "Bold lead:** the report follows"
-
-    Args:
-        prose_line: One line of raw reply text.
-
-    Returns:
-        The line with every leading markdown marker and its padding removed.
-    """
-    return MARKDOWN_LEAD_MARKER_PATTERN.sub("", prose_line.strip(), count=1)
-
-
-def is_list_item_line(prose_line: str) -> bool:
-    """Return True when a line is a bullet or numbered list item.
-
-    Args:
-        prose_line: One non-empty line of reply prose.
-
-    Returns:
-        True when the line opens with a bullet or numbered-step marker.
-    """
-    if BULLET_LINE_PATTERN.match(prose_line):
-        return True
-    return NUMBERED_STEP_PATTERN.match(prose_line) is not None
-
-
-def count_overpacked_list_lines(all_prose_lines: list[str]) -> int:
-    """Return how many list lines carry more words than one idea needs.
-
-    Plain paragraphs are not counted so a requested full report can use long
-    prose lines without a word ceiling.
-
-    Args:
-        all_prose_lines: The non-empty lines of reply prose.
-
-    Returns:
-        The count of bullet/numbered lines over the per-list-line word cap.
-    """
-    return sum(
-        1
-        for each_line in all_prose_lines
-        if is_list_item_line(each_line)
-        and count_words_in_line(each_line) > MAXIMUM_WORDS_PER_LIST_LINE
-    )
-
-
-def describe_bullet_cap_violation(bullet_line_count: int) -> str:
-    """Return the violation text naming the bullet count and the target."""
-    return (
-        f"{bullet_line_count} bullets, over the {MAXIMUM_BULLET_LINE_COUNT}-bullet "
-        f"cap - put findings in at most {TARGET_BULLET_LINE_COUNT} bullets"
-    )
-
-
-def describe_overpacked_list_line_violation(overpacked_line_count: int) -> str:
-    """Return the violation text for overpacked list lines."""
-    return (
-        f"{overpacked_line_count} list lines carry too many words - one idea per "
-        f"bullet or step, at most {MAXIMUM_WORDS_PER_LIST_LINE} words each"
-    )
-
-
-def describe_action_first_violation() -> str:
-    """Return the violation text for instructions buried under prose."""
-    return (
-        "the reply tells the user to act but leads with prose - open with "
-        "numbered steps, one short line each"
-    )
-
-
-def is_action_first_violation(all_prose_lines: list[str]) -> bool:
-    """Return True when multi-line instructions lack leading numbered steps.
-
-    ::
-
-        ok:   single-line "Run the migration." -> False
-        ok:   numbered steps first, then prose -> False
-        flag: prose then "Install the package." -> True
-
-    Args:
-        all_prose_lines: The non-empty lines of reply prose.
-
-    Returns:
-        True when the reply has more than one line, carries instruction
-        lines, and shows no numbered step among its lead lines.
-    """
-    if len(all_prose_lines) <= 1:
-        return False
-    if not has_imperative_instruction_line(all_prose_lines):
-        return False
-    return not has_leading_numbered_step(all_prose_lines)
-
-
-def collect_shape_violations(all_prose_lines: list[str]) -> list[str]:
-    """Return the violation texts for every reply-shape rule the reply breaks.
-
-    Args:
-        all_prose_lines: The non-empty lines of reply prose.
-
-    Returns:
-        One violation text per broken rule, empty when the reply is in shape.
-    """
-    all_violations: list[str] = []
-    if is_action_first_violation(all_prose_lines):
-        all_violations.append(describe_action_first_violation())
-    bullet_line_count = count_bullet_lines(all_prose_lines)
-    if bullet_line_count > MAXIMUM_BULLET_LINE_COUNT:
-        all_violations.append(describe_bullet_cap_violation(bullet_line_count))
-    overpacked_list_line_count = count_overpacked_list_lines(all_prose_lines)
-    if overpacked_list_line_count > MAXIMUM_OVERPACKED_LIST_LINE_COUNT:
-        all_violations.append(
-            describe_overpacked_list_line_violation(overpacked_list_line_count)
-        )
-    return all_violations
-
-
-def find_reply_shape_violations(assistant_message: str) -> list[str]:
-    """Return every reply-shape violation the final assistant message carries.
-
-    ::
-
-        ok:   a short outcome sentence              -> []
-        ok:   a long requested audit with few bullets -> []
-        flag: prose then buried "Install the package." -> [action-first]
-
-    Args:
-        assistant_message: The raw final assistant message.
-
-    Returns:
-        One violation text per broken rule, empty when the reply is in shape.
-    """
-    prose_text = extract_reply_prose(assistant_message)
-    return collect_shape_violations(collect_non_empty_lines(prose_text))
-
-
-def build_block_reason(all_violations: list[str]) -> str:
-    """Return the corrective message naming each violation and the rewrite path.
-
-    Args:
-        all_violations: The violation texts the reply earned.
-
-    Returns:
-        The full block reason the model rewrites its reply against.
-    """
-    formatted_violation_list = VIOLATION_SEPARATOR.join(all_violations)
-    return (
-        f"ELI11 REPLY SHAPE: Rewrite the reply into a short action-first shape "
-        f"({formatted_violation_list}).\n\n"
-        f"When the user must act, open with numbered click-by-click steps, one "
-        f"short line each. When nothing is needed, open with the outcome in one "
-        f"sentence. Keep findings to at most {TARGET_BULLET_LINE_COUNT} short "
-        f"bullets. A requested full report may run long; keep list lines short "
-        f"and lead with the outcome. Code fences, blockquotes, tables, and "
-        f"links are already exempt from shape counting.\n\n"
-        f"You MUST re-output the complete, revised response with the correction "
-        f"applied."
-    )
-
-
-def main() -> None:
-    """Read the Stop payload and block a reply that breaks the ELI11 shape."""
-    try:
-        hook_input = json.load(sys.stdin)
-    except json.JSONDecodeError:
-        sys.exit(0)
-
-    if hook_input.get("stop_hook_active", False):
-        sys.exit(0)
-
-    assistant_message = hook_input.get("last_assistant_message", "")
-
-    if not assistant_message:
-        sys.exit(0)
-
-    all_violations = find_reply_shape_violations(assistant_message)
-
-    if not all_violations:
-        sys.exit(0)
-
-    block_reason = build_block_reason(all_violations)
-    block_response = {
-        "decision": "block",
-        "reason": block_reason,
-        "systemMessage": USER_FACING_ELI11_NOTICE,
-        "suppressOutput": True,
-    }
-    log_hook_block(
-        calling_hook_name="eli11_reply_enforcer.py",
-        hook_event="Stop",
-        block_reason=block_reason,
-    )
-    print(json.dumps(block_response))
-    sys.exit(0)
-
-
-if __name__ == "__main__":
-    main()