sourcecode 1.30.27__tar.gz → 1.30.28__tar.gz

{sourcecode-1.30.27 → sourcecode-1.30.28}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: sourcecode
-Version: 1.30.27
+Version: 1.30.28
 Summary: Deterministic codebase context for AI coding agents
 License:                                  Apache License
                                    Version 2.0, January 2004
@@ -221,7 +221,7 @@ Description-Content-Type: text/markdown
 
 **Deterministic, behavior-aware codebase context for AI agents and PR review.**
 
-![Version](https://img.shields.io/badge/version-1.30.27-blue)
+![Version](https://img.shields.io/badge/version-1.30.28-blue)
 ![Python](https://img.shields.io/badge/python-3.10%2B-green)
 
 ---
@@ -257,7 +257,7 @@ pipx install sourcecode
 
 ```bash
 sourcecode version
-# sourcecode 1.30.27
+# sourcecode 1.30.28
 ```
 
 ---

{sourcecode-1.30.27 → sourcecode-1.30.28}/README.md

@@ -2,7 +2,7 @@
 
 **Deterministic, behavior-aware codebase context for AI agents and PR review.**
 
-![Version](https://img.shields.io/badge/version-1.30.27-blue)
+![Version](https://img.shields.io/badge/version-1.30.28-blue)
 ![Python](https://img.shields.io/badge/python-3.10%2B-green)
 
 ---
@@ -38,7 +38,7 @@ pipx install sourcecode
 
 ```bash
 sourcecode version
-# sourcecode 1.30.27
+# sourcecode 1.30.28
 ```
 
 ---

{sourcecode-1.30.27 → sourcecode-1.30.28}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "hatchling.build"
 
 [project]
 name = "sourcecode"
-version = "1.30.27"
+version = "1.30.28"
 description = "Deterministic codebase context for AI coding agents"
 readme = "README.md"
 requires-python = ">=3.9"

{sourcecode-1.30.27 → sourcecode-1.30.28}/src/sourcecode/__init__.py

@@ -1,3 +1,3 @@
 """sourcecode — Deterministic codebase context maps for AI coding agents."""
 
-__version__ = "1.30.27"
+__version__ = "1.30.28"

{sourcecode-1.30.27 → sourcecode-1.30.28}/src/sourcecode/cli.py

@@ -424,7 +424,7 @@ def main(
     no_redact: bool = typer.Option(
         False,
         "--no-redact",
-        help="Disable secret redaction. Use with caution — output may contain sensitive values.",
+        help="Disable secret redaction of output strings. Note: env var values from the OS are never included in output regardless of this flag (security policy).",
     ),
     version: Optional[bool] = typer.Option(
         None,
@@ -437,7 +437,7 @@ def main(
     depth: int = typer.Option(
         4,
         "--depth",
-        help="File tree traversal depth (default: 4). Java/Maven projects auto-adjust to 12.",
+        help="File tree traversal depth (default: 4). Java/Maven projects auto-adjust to a minimum of 12; values below 12 have no effect on Java projects.",
         min=1,
         max=20,
     ),
@@ -1734,6 +1734,11 @@ def prepare_context_cmd(
         help="Emit per-phase timing to stderr (git scan ms, symptom scoring ms, total ms)",
         hidden=True,
     ),
+    fast: bool = typer.Option(
+        False,
+        "--fast",
+        help="Skip deep analysis (content search, test gap discovery, code annotations). Uses manifest/metadata only. Target: < 6 s.",
+    ),
 ) -> None:
     """Task-specific context for AI coding agents.
 
@@ -1808,9 +1813,14 @@ def prepare_context_cmd(
     if since:
         _phase += f" since {since}"
     _progress.start(_phase)
+    if not fast:
+        import sys as _sys
+        if _sys.stderr.isatty():
+            _sys.stderr.write(f"Analyzing ({task})... (deep scan may take 15–35 s for large codebases)\n")
+            _sys.stderr.flush()
     _t0 = _time.perf_counter()
     try:
-        output = builder.build(task, since=since, symptom=symptom)
+        output = builder.build(task, since=since, symptom=symptom, fast=fast)
     finally:
         _progress.finish()
     _t_total = (_time.perf_counter() - _t0) * 1000
@@ -1960,8 +1970,8 @@ def prepare_context_cmd(
                 _sys.stdout.buffer.write(b"\n")
                 _sys.stdout.buffer.flush()
             if copy:
-                _copy_to_clipboard(_nc_json)
-                typer.echo("✓ copied to clipboard", err=True)
+                if _copy_to_clipboard(_nc_json):
+                    typer.echo("✓ copied to clipboard", err=True)
             raise typer.Exit()
         if output.ci_decision:
             out["ci_decision"] = output.ci_decision
@@ -2082,6 +2092,10 @@ def prepare_context_cmd(
         out["symptom_note"] = output.symptom_note
     if output.symptom_explain:
         out["symptom_explain"] = output.symptom_explain
+    if getattr(output, "symptom_hint", None):
+        out["symptom_hint"] = output.symptom_hint
+    if getattr(output, "warnings", None):
+        out["warnings"] = output.warnings
     if llm_prompt:
         out["llm_prompt"] = builder.render_prompt(output)
 

{sourcecode-1.30.27 → sourcecode-1.30.28}/src/sourcecode/prepare_context.py

@@ -335,6 +335,7 @@ class TaskOutput:
     related_notes: list[dict] = field(default_factory=list)        # fix-bug + symptom only
     symptom_note: Optional[str] = None                             # fix-bug: cross-layer synonym note
     symptom_explain: Optional[dict] = None                         # fix-bug: structured evidence breakdown
+    symptom_hint: Optional[str] = None                             # fix-bug: redirect hint when term not found in this module
     # delta-specific impact fields
     impact_summary: Optional[str] = None
     affected_modules: list[str] = field(default_factory=list)
@@ -348,6 +349,7 @@ class TaskOutput:
     error_code: Optional[str] = None
     error_message: Optional[str] = None
     error_hints: list[str] = field(default_factory=list)
+    warnings: list[dict] = field(default_factory=list)             # structured warnings (REF_NOT_FOUND, etc.)
     # CI decision state machine — machine-decidable signal
     ci_decision: Optional[str] = None  # "no_changes" | "analysis_success" | "git_ref_error" | "no_git_repo"
     # git baseline resolution metadata
@@ -683,7 +685,7 @@ class TaskContextBuilder:
     def __init__(self, root: Path) -> None:
         self.root = root
 
-    def build(self, task_name: str, *, since: Optional[str] = None, symptom: Optional[str] = None) -> TaskOutput:
+    def build(self, task_name: str, *, since: Optional[str] = None, symptom: Optional[str] = None, fast: bool = False) -> TaskOutput:
         if task_name not in TASKS:
             raise ValueError(
                 f"Unknown task '{task_name}'. Available: {', '.join(TASKS)}"
@@ -880,7 +882,7 @@ class TaskContextBuilder:
         improvement_opportunities: list[str] = []
         cn_notes_for_ranking: list = []
 
-        if spec.enable_code_notes:
+        if spec.enable_code_notes and not fast:
             from dataclasses import asdict
             from sourcecode.code_notes_analyzer import CodeNotesAnalyzer
 
@@ -1317,6 +1319,7 @@ class TaskContextBuilder:
         related_notes: list[dict] = []
         symptom_note: Optional[str] = None
         symptom_explain: Optional[dict] = None
+        symptom_hint: Optional[str] = None
         if task_name == "fix-bug" and symptom:
             import re as _re
             _camel_expanded = _re.sub(r'([a-z])([A-Z])', r'\1 \2', symptom)
@@ -1537,9 +1540,30 @@ class TaskContextBuilder:
                     ),
                 }
 
+                # BUG #4: LOW confidence + 0 content matches → clear suspected_areas,
+                # emit actionable redirect instead of unrelated files.
+                if _sx_confidence == "LOW" and not _sx_content:
+                    suspected_areas = []
+                    _is_fe_term = any(kw in _FRONTEND_SYMPTOM_MAP for kw in symptom_keywords)
+                    _root_name = self.root.name
+                    if _is_fe_term:
+                        _fe_redirect = (
+                            f"Term {symptom!r} not found in sources under {_root_name!r}. "
+                            f"This appears to be a frontend symptom. "
+                            f"Try: prepare-context fix-bug . --symptom {symptom!r} "
+                            f"(monorepo root) or target a frontend sub-project directly."
+                        )
+                    else:
+                        _fe_redirect = (
+                            f"Term {symptom!r} not found in sources under {_root_name!r}. "
+                            f"Verify the spelling or try a related term. "
+                            f"If this is a frontend symptom, run against the frontend sub-project."
+                        )
+                    symptom_hint = _fe_redirect
+
         # ── 7. Test gaps (generate-tests only) ────────────────────────────
         test_gaps: list[str] = []
-        if task_name == "generate-tests":
+        if task_name == "generate-tests" and not fast:
             def _normalize_test_stem(stem: str) -> str:
                 # Java: FooTest / FooTests → Foo; TestFoo → Foo
                 if stem.endswith("Tests"):
@@ -1683,6 +1707,8 @@ class TaskContextBuilder:
             resolved_since_ref=_delta_baseline.get("resolved_ref") if task_name == "delta" else None,
             resolution_path=_delta_baseline.get("resolution_path") if task_name == "delta" else None,
             diff_validation_status=_delta_baseline.get("diff_validation_status") if task_name == "delta" else None,
+            warnings=_delta_baseline.get("warnings", []) if task_name == "delta" else [],
+            symptom_hint=symptom_hint if task_name == "fix-bug" else None,
         )
 
     def render_prompt(self, output: TaskOutput) -> str:
@@ -3261,6 +3287,7 @@ class TaskContextBuilder:
                         "resolution_path": "head_minus_1_fallback",
                         "diff_validation_status": "invalid_ref",  # original ref unresolved
                         "error": False,
+                        "warnings": [{"code": "REF_NOT_FOUND", "ref": since, "resolved_to": "HEAD~1"}],
                     }
 
             # All stages failed

{sourcecode-1.30.27 → sourcecode-1.30.28}/src/sourcecode/serializer.py

@@ -248,6 +248,16 @@ def _compact_git_context(sm: "SourceMap") -> "Optional[dict[str, Any]]":
                 for f, n in _fc.most_common(5)
             ]
             ctx["hotspots_source"] = "recent_commits"
+    if gc.recent_commits:
+        ctx["recent_commits"] = [
+            {
+                "hash": c.hash[:8],
+                "message": (c.message or "")[:80],
+                "date": (c.date or "")[:10],
+                "author": c.author or "",
+            }
+            for c in gc.recent_commits[:5]
+        ]
     return ctx if ctx else None
 
 

sourcecode-1.30.28/tests/test_bug_fixes_v1302.py

@@ -0,0 +1,419 @@
+"""Tests for v1.30.27 bug fixes and improvements.
+
+  BUG #1  --copy confirmation goes to stderr (not stdout)
+  BUG #2  --since non-existent branch emits warnings[] in output
+  BUG #3  --depth help text documents Java auto-adjust behavior
+  BUG #4  symptom LOW-confidence: suspected_areas cleared, symptom_hint present
+  IMP #5  --fast flag skips deep analysis; progress message gated on TTY
+  IMP #6  git_context compact output includes recent_commits
+  IMP #7  --no-redact help text documents security policy
+"""
+from __future__ import annotations
+
+import json
+from pathlib import Path
+from typing import Any
+from unittest.mock import patch, MagicMock
+
+import pytest
+from typer.testing import CliRunner
+
+from sourcecode.cli import app, _copy_to_clipboard
+from sourcecode import prepare_context as _pc
+from sourcecode.prepare_context import TaskContextBuilder, TaskOutput
+from sourcecode.schema import GitContext, CommitRecord, ChangeHotspot, UncommittedChanges, SourceMap, AnalysisMetadata
+from sourcecode.serializer import _compact_git_context
+
+runner = CliRunner()
+
+FIXTURE = Path(__file__).parent / "fixtures" / "fastapi_app"
+
+
+def _invoke(*args: str):
+    return runner.invoke(app, list(args))
+
+
+def _json(result) -> dict:
+    try:
+        return json.loads(result.output)
+    except json.JSONDecodeError as e:
+        raise AssertionError(f"Not valid JSON. output={result.output!r}") from e
+
+
+# ── BUG #1: --copy confirmation must go to stderr ────────────────────────────
+
+class TestCopyToStderr:
+    def test_copy_confirmation_uses_err_flag(self, tmp_path):
+        """typer.echo for clipboard confirmation uses err=True (routes to stderr).
+
+        CliRunner mixes stderr into result.output, so we verify the JSON is
+        parseable (as the leading content) and the message appears after the JSON.
+        In real piped usage, stdout only carries JSON; stderr carries the message.
+        """
+        (tmp_path / "main.py").write_text("x = 1\n")
+        with patch("sourcecode.cli._copy_to_clipboard", return_value=True):
+            result = runner.invoke(app, [str(tmp_path), "--compact", "--copy"])
+        assert result.exit_code == 0
+        # The JSON portion at the start must be parseable via raw_decode
+        decoder = json.JSONDecoder()
+        data, end_idx = decoder.raw_decode(result.output.lstrip())
+        assert isinstance(data, dict)
+        # Confirmation must appear AFTER the JSON (proves it was stderr, not inline)
+        trailing = result.output[end_idx:].strip()
+        assert "copied to clipboard" in trailing, (
+            "Confirmation must be in stderr (appears after JSON in CliRunner output)"
+        )
+
+    def test_copy_failure_no_message(self, tmp_path):
+        """No confirmation when clipboard copy fails."""
+        (tmp_path / "main.py").write_text("x = 1\n")
+        with patch("sourcecode.cli._copy_to_clipboard", return_value=False):
+            result = runner.invoke(app, [str(tmp_path), "--compact", "--copy"])
+        assert result.exit_code == 0
+        assert "copied to clipboard" not in result.output
+
+    def test_prepare_context_copy_conditional_on_success(self, tmp_path):
+        """prepare-context --copy message only shown when clipboard succeeds."""
+        (tmp_path / "main.py").write_text("x = 1\n")
+        with patch("sourcecode.cli._copy_to_clipboard", return_value=False):
+            result = runner.invoke(
+                app, ["prepare-context", "explain", str(tmp_path), "--copy"]
+            )
+        # No "copied to clipboard" in any output when copy fails
+        assert "copied to clipboard" not in result.output
+
+
+# ── BUG #2: --since non-existent ref emits warnings[] ────────────────────────
+
+class TestSinceWarnings:
+    def test_head_minus1_fallback_emits_warning(self, tmp_path):
+        """Stage 4 fallback populates warnings[] with REF_NOT_FOUND."""
+        builder = TaskContextBuilder(tmp_path)
+
+        # Simulate Stage 4: ref invalid, HEAD~1 available
+        def _mock_resolve(since):
+            return {
+                "files": ["main.py"],
+                "resolved_ref": "HEAD~1",
+                "resolution_path": "head_minus_1_fallback",
+                "diff_validation_status": "invalid_ref",
+                "error": False,
+                "warnings": [{"code": "REF_NOT_FOUND", "ref": since, "resolved_to": "HEAD~1"}],
+            }
+
+        (tmp_path / "main.py").write_text("x = 1\n")
+        with patch.object(builder, "_resolve_git_baseline", side_effect=_mock_resolve):
+            with patch.object(builder, "_build_delta_impact") as mock_impact:
+                mock_impact.return_value = MagicMock(
+                    relevant_files=[], impact_summary=None, affected_modules=[],
+                    risk_areas=[], why_these_files={}, analysis_gaps=[],
+                    system_impact={}, change_type=[], dependency_graph_summary={},
+                    impact_score_per_file={},
+                )
+                with patch.object(builder, "_get_pr_scope_files", return_value=(None, "git_diff", [], [])):
+                    output = builder.build("delta", since="main")
+
+        assert output.warnings, "warnings must be non-empty for Stage 4 fallback"
+        assert output.warnings[0]["code"] == "REF_NOT_FOUND"
+        assert output.warnings[0]["ref"] == "main"
+        assert output.warnings[0]["resolved_to"] == "HEAD~1"
+
+    def test_valid_ref_no_warnings(self, tmp_path):
+        """Exact ref resolution (Stage 1) produces no warnings."""
+        builder = TaskContextBuilder(tmp_path)
+
+        def _mock_resolve(since):
+            return {
+                "files": [],
+                "resolved_ref": since,
+                "resolution_path": "exact_local_ref",
+                "diff_validation_status": "valid_empty",
+                "error": False,
+            }
+
+        with patch.object(builder, "_resolve_git_baseline", side_effect=_mock_resolve):
+            with patch.object(builder, "_build_delta_impact") as mock_impact:
+                mock_impact.return_value = MagicMock(
+                    relevant_files=[], impact_summary=None, affected_modules=[],
+                    risk_areas=[], why_these_files={}, analysis_gaps=[],
+                    system_impact={}, change_type=[], dependency_graph_summary={},
+                    impact_score_per_file={},
+                )
+                output = builder.build("delta", since="origin/develop")
+
+        assert not output.warnings
+
+    def test_warnings_serialized_in_cli_output(self, tmp_path):
+        """warnings[] appears in JSON output when populated."""
+        (tmp_path / "main.py").write_text("x = 1\n")
+
+        mock_output = MagicMock(spec=TaskOutput)
+        mock_output.task = "delta"
+        mock_output.goal = "g"
+        mock_output.project_summary = None
+        mock_output.architecture_summary = None
+        mock_output.relevant_files = []
+        mock_output.suspected_areas = []
+        mock_output.improvement_opportunities = []
+        mock_output.test_gaps = []
+        mock_output.key_dependencies = []
+        mock_output.code_notes_summary = None
+        mock_output.limitations = []
+        mock_output.confidence = "low"
+        mock_output.gaps = []
+        mock_output.why_these_files = {}
+        mock_output.changed_files = []
+        mock_output.affected_entry_points = []
+        mock_output.symptom = None
+        mock_output.related_notes = []
+        mock_output.symptom_note = None
+        mock_output.symptom_explain = None
+        mock_output.symptom_hint = None
+        mock_output.impact_summary = None
+        mock_output.affected_modules = []
+        mock_output.risk_areas = []
+        mock_output.since = "main"
+        mock_output.system_impact = {}
+        mock_output.change_type = []
+        mock_output.dependency_graph_summary = {}
+        mock_output.impact_score_per_file = {}
+        mock_output.error_code = None
+        mock_output.error_message = None
+        mock_output.error_hints = []
+        mock_output.ci_decision = "analysis_success"
+        mock_output.resolved_since_ref = "HEAD~1"
+        mock_output.resolution_path = "head_minus_1_fallback"
+        mock_output.diff_validation_status = "invalid_ref"
+        mock_output.warnings = [{"code": "REF_NOT_FOUND", "ref": "main", "resolved_to": "HEAD~1"}]
+        mock_output.base_ref = None
+        mock_output.security_impact = {}
+        mock_output.transactional_impact = {}
+        mock_output.configuration_impact = {}
+        mock_output.test_coverage_risk = {}
+        mock_output.review_hotspots = []
+        mock_output.suggested_review_order = []
+        mock_output.execution_paths = []
+        mock_output.behavioral_impact = []
+        mock_output.scope_source = None
+        mock_output.scope_files = []
+        mock_output.repo_root = None
+        mock_output.runtime_changes = []
+        mock_output.build_changes = {}
+        mock_output.committed_changes = []
+        mock_output.uncommitted_changes = []
+        mock_output.analysis_scope = {}
+
+        with patch.object(TaskContextBuilder, "build", return_value=mock_output):
+            result = runner.invoke(
+                app, ["prepare-context", "delta", str(tmp_path), "--since", "main"]
+            )
+
+        assert result.exit_code == 0
+        data = _json(result)
+        assert "warnings" in data
+        assert data["warnings"][0]["code"] == "REF_NOT_FOUND"
+
+
+# ── BUG #3: --depth help text ─────────────────────────────────────────────────
+
+class TestDepthHelp:
+    def test_depth_help_documents_java_autoadjust(self):
+        result = runner.invoke(app, ["--help"])
+        assert result.exit_code == 0
+        assert "12" in result.output  # Java min depth mentioned
+        # The help text should mention auto-adjust or Java
+        assert "Java" in result.output or "auto" in result.output.lower()
+
+    def test_depth_flag_accepted(self, tmp_path):
+        (tmp_path / "main.py").write_text("x = 1\n")
+        result = runner.invoke(app, [str(tmp_path), "--depth", "6", "--compact"])
+        assert result.exit_code == 0
+
+
+# ── BUG #4: symptom LOW-confidence clears suspected_areas, adds symptom_hint ──
+
+class TestSymptomHint:
+    def _make_java_fixture(self, tmp_path: Path) -> Path:
+        src = tmp_path / "src" / "main" / "java" / "com" / "example"
+        src.mkdir(parents=True)
+        (src / "UserService.java").write_text(
+            "package com.example;\npublic class UserService {}\n"
+        )
+        (tmp_path / "pom.xml").write_text(
+            "<project><groupId>com.example</groupId></project>\n"
+        )
+        return tmp_path
+
+    def test_frontend_symptom_in_java_module_clears_suspected_areas(self, tmp_path):
+        """'spinner' in Java module → suspected_areas=[], symptom_hint present."""
+        self._make_java_fixture(tmp_path)
+        result = runner.invoke(
+            app,
+            ["prepare-context", "fix-bug", str(tmp_path), "--symptom", "spinner"],
+        )
+        assert result.exit_code == 0
+        data = _json(result)
+        # suspected_areas must be empty when confidence is LOW with 0 content matches
+        if data.get("symptom_explain", {}).get("confidence") == "LOW":
+            assert data.get("suspected_areas", []) == [], (
+                "suspected_areas must be [] when symptom confidence is LOW"
+            )
+            assert "symptom_hint" in data, "symptom_hint must be present when LOW confidence"
+            assert "spinner" in data["symptom_hint"].lower() or "frontend" in data["symptom_hint"].lower()
+
+    def test_java_symptom_no_hint(self, tmp_path):
+        """Java term in Java module → no symptom_hint injected."""
+        self._make_java_fixture(tmp_path)
+        result = runner.invoke(
+            app,
+            ["prepare-context", "fix-bug", str(tmp_path), "--symptom", "UserService"],
+        )
+        assert result.exit_code == 0
+        data = _json(result)
+        # If confidence is not LOW, no symptom_hint should be present
+        sx = data.get("symptom_explain", {})
+        if sx.get("confidence") != "LOW":
+            assert "symptom_hint" not in data
+
+    def test_symptom_hint_field_on_taskoutput(self):
+        """TaskOutput dataclass has symptom_hint field."""
+        out = TaskOutput(
+            task="fix-bug", goal="g", project_summary=None,
+            architecture_summary=None, relevant_files=[], suspected_areas=[],
+            improvement_opportunities=[], test_gaps=[], key_dependencies=[],
+            code_notes_summary=None, limitations=[], symptom_hint="redirect hint",
+        )
+        assert out.symptom_hint == "redirect hint"
+
+
+# ── IMPROVEMENT #5: --fast flag ───────────────────────────────────────────────
+
+class TestFastFlag:
+    def test_fast_flag_accepted(self, tmp_path):
+        """--fast flag accepted without error."""
+        (tmp_path / "main.py").write_text("x = 1\n")
+        result = runner.invoke(
+            app, ["prepare-context", "explain", str(tmp_path), "--fast"]
+        )
+        assert result.exit_code == 0
+        data = _json(result)
+        assert "task" in data
+
+    def test_fast_skips_code_notes(self, tmp_path):
+        """--fast produces output without triggering code notes analysis."""
+        (tmp_path / "main.py").write_text("# TODO: fix this\nx = 1\n")
+        (tmp_path / "pyproject.toml").write_text('[project]\nname = "test"\n')
+
+        called = []
+        orig_analyze = None
+        try:
+            from sourcecode.code_notes_analyzer import CodeNotesAnalyzer
+            orig_analyze = CodeNotesAnalyzer.analyze
+
+            def _mock_analyze(self, root):
+                called.append(root)
+                return orig_analyze(self, root)
+
+            with patch.object(CodeNotesAnalyzer, "analyze", _mock_analyze):
+                result = runner.invoke(
+                    app, ["prepare-context", "fix-bug", str(tmp_path), "--fast"]
+                )
+        except ImportError:
+            pytest.skip("CodeNotesAnalyzer not importable")
+
+        assert result.exit_code == 0
+        # With --fast, code notes analyzer should NOT have been called
+        assert not called, "--fast must skip code notes analysis"
+
+    def test_fast_flag_in_help(self):
+        """--fast appears in prepare-context help."""
+        result = runner.invoke(app, ["prepare-context", "--help"])
+        assert result.exit_code == 0
+        assert "--fast" in result.output
+
+    def test_no_fast_does_not_contaminate_stdout(self, tmp_path):
+        """Progress message (when not --fast) must not appear in JSON stdout."""
+        (tmp_path / "main.py").write_text("x = 1\n")
+        result = runner.invoke(
+            app, ["prepare-context", "explain", str(tmp_path)]
+        )
+        assert result.exit_code == 0
+        # stdout must parse as valid JSON
+        data = _json(result)
+        assert "task" in data
+
+
+# ── IMPROVEMENT #6: git_context recent_commits ───────────────────────────────
+
+class TestRecentCommitsInGitContext:
+    def _make_sm_with_commits(self, commits: list) -> SourceMap:
+        sm = SourceMap(metadata=AnalysisMetadata(analyzed_path="/tmp/test"))
+        sm.git_context = GitContext(
+            requested=True,
+            branch="main",
+            recent_commits=commits,
+            change_hotspots=[ChangeHotspot(file="a.py", commit_count=3, last_changed="2026-05-01")],
+            uncommitted_changes=UncommittedChanges(staged=[], unstaged=[], untracked=[]),
+        )
+        return sm
+
+    def test_recent_commits_present_in_compact_output(self):
+        """_compact_git_context includes recent_commits with top-5 entries."""
+        commits = [
+            CommitRecord(hash="abcd1234", message="feat: add user endpoint", author="alice", date="2026-05-15", files_changed=["a.py"]),
+            CommitRecord(hash="efgh5678", message="fix: null pointer in service", author="bob", date="2026-05-14", files_changed=["b.py"]),
+        ]
+        sm = self._make_sm_with_commits(commits)
+        ctx = _compact_git_context(sm)
+        assert ctx is not None
+        assert "recent_commits" in ctx
+        assert len(ctx["recent_commits"]) == 2
+
+    def test_recent_commits_fields(self):
+        """Each recent_commits entry has hash, message, date, author."""
+        commits = [
+            CommitRecord(
+                hash="a0b438a4deadbeef",
+                message="Merge pull request #82 into main — very long message that should be truncated at exactly eighty characters total",
+                author="m3-dhl",
+                date="2026-05-15",
+                files_changed=[],
+            ),
+        ]
+        sm = self._make_sm_with_commits(commits)
+        ctx = _compact_git_context(sm)
+        entry = ctx["recent_commits"][0]
+        assert entry["hash"] == "a0b438a4"   # 8-char short hash
+        assert len(entry["message"]) <= 80
+        assert entry["date"] == "2026-05-15"
+        assert entry["author"] == "m3-dhl"
+
+    def test_recent_commits_capped_at_5(self):
+        """At most 5 commits in recent_commits."""
+        commits = [
+            CommitRecord(hash=f"hash{i:08d}", message=f"commit {i}", author="a", date="2026-05-01", files_changed=[])
+            for i in range(10)
+        ]
+        sm = self._make_sm_with_commits(commits)
+        ctx = _compact_git_context(sm)
+        assert len(ctx["recent_commits"]) == 5
+
+    def test_no_commits_no_key(self):
+        """No recent_commits key when list is empty."""
+        sm = self._make_sm_with_commits([])
+        ctx = _compact_git_context(sm)
+        assert ctx is None or "recent_commits" not in (ctx or {})
+
+
+# ── IMPROVEMENT #7: --no-redact help text ────────────────────────────────────
+
+class TestNoRedactHelp:
+    def test_no_redact_help_mentions_env_var_policy(self):
+        """--no-redact help text must clarify env var values are never included."""
+        result = runner.invoke(app, ["--help"])
+        assert result.exit_code == 0
+        assert "--no-redact" in result.output
+        # The help must not imply env var values will be shown
+        # Verify the security policy note is present somewhere in help
+        # (exact wording may vary — test for the key concepts)
+        assert "security" in result.output.lower() or "policy" in result.output.lower() or "never" in result.output.lower()