agent2learn 0.1.2__tar.gz → 0.1.6__tar.gz

{agent2learn-0.1.2 → agent2learn-0.1.6}/PKG-INFO

@@ -1,12 +1,12 @@
 Metadata-Version: 2.4
 Name: agent2learn
-Version: 0.1.2
+Version: 0.1.6
 Summary: A local, source-cited course vault for AI agents
 License-Expression: Apache-2.0
 License-File: LICENSE
-Requires-Dist: typer>=0.15,<1
+Requires-Dist: typer>=0.16,<1
 Requires-Dist: click>=8.2,<9
-Requires-Dist: rich>=13,<15
+Requires-Dist: rich>=13,<16
 Requires-Dist: requests>=2.32,<3
 Requires-Dist: platformdirs>=4.2,<5
 Requires-Dist: keyring>=25,<27
@@ -34,19 +34,19 @@ a durable local vault: original files beside Markdown twins, a stable index, and
 resolve to ordinary paths and line numbers. Your agent stops guessing about your course and starts
 quoting it.
 
-> **Walkthrough recording:** not yet attached. It is a release artifact, recorded against the
-> synthetic fixture course in `tests/fixtures/` — never against a real account — so no course,
-> student, grade, or session data appears in it. Recording it is a launch gate in
-> [docs/LAUNCH.md](docs/LAUNCH.md); this README will not link a file that does not exist.
-
 ## Install
 
-Three supported options. Pick one.
+For the measured macOS workflow, repaired defects, PDF coverage gaps, and still-unverified
+human/platform checks, see the [release validation scope](https://github.com/ManagementMO/agent2learn/blob/main/docs/RELEASE_VALIDATION.md).
+Automated cross-platform tests are not a claim that every live login or course conversion works.
+
+For most students, use the command for your operating system. In a real terminal it installs
+Agent2Learn and starts guided setup immediately. If you already have uv, use the uv option below.
 
 **macOS and Linux**
 
 ```bash
-curl -fsSL https://raw.githubusercontent.com/ManagementMO/agent2learn/main/install.sh | bash
+(installer="$(curl -fsSL https://raw.githubusercontent.com/ManagementMO/agent2learn/main/install.sh)" && bash -c "$installer")
 ```
 
 **Windows (PowerShell)**
@@ -57,21 +57,35 @@ irm https://raw.githubusercontent.com/ManagementMO/agent2learn/main/install.ps1
 
 **Already have [uv](https://docs.astral.sh/uv/)?**
 
+Bash or zsh:
+
 ```bash
-uv tool install agent2learn
-a2l init
+uv tool install agent2learn && uv tool run --from agent2learn a2l init
+```
+
+Windows PowerShell:
+
+```powershell
+uv tool install agent2learn; if ($LASTEXITCODE -eq 0) { uv tool run --from agent2learn a2l init } else { throw "Agent2Learn installation failed" }
 ```
 
+These uv commands start setup through uv itself, so a newly installed `a2l` does not need to be
+on the current shell's PATH first. `uv tool install agent2learn` remains the install-only command.
+
 These commands install the published package from PyPI, not an unreleased repository checkout.
 Python 3.11–3.14 is supported. If uv selects an older system Python or your shell cannot find
-`a2l`, see [setup troubleshooting](docs/FAQ.md#setup-and-path).
+`a2l`, see [setup troubleshooting](https://github.com/ManagementMO/agent2learn/blob/main/docs/FAQ.md#setup-and-path).
 
-`install.sh` and `install.ps1` install a pinned Agent2Learn (currently 0.1.2), verify that `a2l`
-runs, and then **continue straight into interactive `a2l init`** in the same command. If they are
-run without a terminal on both ends — in CI, or through a pipe — they stop after verifying and
-print the exact next step instead of setting anything up. Neither script needs administrator
-rights, and neither creates a vault, installs agent skills, or opens a browser by itself.
-Onboarding does those things, after showing you a preview and asking.
+`install.sh` and `install.ps1` install a pinned Agent2Learn release, verify that `a2l` runs,
+and then **continue straight into interactive `a2l init`** in the same command. The
+macOS/Linux launcher waits for a successful download and preserves your terminal for setup prompts.
+If a script runs without a terminal on both ends, such as in CI or a headless agent tool, it stops
+after verification and prints the next step instead of setting anything up.
+
+Neither script needs administrator rights or creates a vault, installs agent skills, or opens a
+browser by itself. Onboarding does those things after showing you a preview and asking; you still
+complete LEARN sign-in and Duo yourself. Already-open terminals or agents may need to be restarted
+to pick up the new executable path, but you do not need to repeat completed setup.
 
 ### Agent skills
 
@@ -83,7 +97,7 @@ Separately, `npx skills add ManagementMO/agent2learn` is an optional skills-only
 skills ecosystem. It **does not install** the `a2l` engine — only the skill documents — so the
 commands they describe will not exist until you install Agent2Learn itself with one of the three
 options above. That route also runs a third-party CLI with its own network behaviour; see
-[docs/PRIVACY.md](docs/PRIVACY.md).
+[docs/PRIVACY.md](https://github.com/ManagementMO/agent2learn/blob/main/docs/PRIVACY.md).
 
 ## Then just ask
 
@@ -146,7 +160,7 @@ a2l submit COURSE ITEM FILE      preview an upload, then require your typed conf
 
 Agent2Learn is **not affiliated with, endorsed by, or supported by the University of Waterloo or
 D2L Corporation**. You are responsible for using it within your course rules and your institution's
-policies. See [DISCLAIMER.md](DISCLAIMER.md).
+policies. See [DISCLAIMER.md](https://github.com/ManagementMO/agent2learn/blob/main/DISCLAIMER.md).
 
 ## Privacy defaults
 
@@ -161,26 +175,28 @@ policies. See [DISCLAIMER.md](DISCLAIMER.md).
 
 Turning collection off does not delete what is already on disk; `a2l privacy purge` does that, and
 it previews the exact targets and requires a typed phrase first. Full detail, including every
-external request Agent2Learn can make, is in [docs/PRIVACY.md](docs/PRIVACY.md).
+external request Agent2Learn can make, is in [docs/PRIVACY.md](https://github.com/ManagementMO/agent2learn/blob/main/docs/PRIVACY.md).
 
 ## Documentation
 
-- [Install guide for agents](docs/install.md) — point your agent here and let it handle setup.
-- [Authentication](docs/AUTHENTICATION.md) — the dedicated profile, Duo, expiry, and recovery.
-- [Privacy](docs/PRIVACY.md) — what is stored, where, every network action, and how to delete it.
-- [FAQ](docs/FAQ.md) — the things that actually go wrong.
-- [Porting to another school](docs/PORTING.md) — the `School` protocol and a worked reference.
-- [What is deferred, and why](docs/FUTURE.md)
-- [Security policy](SECURITY.md) · [Disclaimer](DISCLAIMER.md) · [Licence](LICENSE) ·
-  [Third-party notices](THIRD_PARTY_NOTICES.md)
+- [Install guide for agents](https://github.com/ManagementMO/agent2learn/blob/main/docs/install.md) — point your agent here and let it handle setup.
+- [Authentication](https://github.com/ManagementMO/agent2learn/blob/main/docs/AUTHENTICATION.md) — the dedicated profile, Duo, expiry, and recovery.
+- [Privacy](https://github.com/ManagementMO/agent2learn/blob/main/docs/PRIVACY.md) — what is stored, where, every network action, and how to delete it.
+- [FAQ](https://github.com/ManagementMO/agent2learn/blob/main/docs/FAQ.md) — the things that actually go wrong.
+- [Porting to another school](https://github.com/ManagementMO/agent2learn/blob/main/docs/PORTING.md) — the `School` protocol and a worked reference.
+- [What is deferred, and why](https://github.com/ManagementMO/agent2learn/blob/main/docs/FUTURE.md)
+- [Security policy](https://github.com/ManagementMO/agent2learn/blob/main/SECURITY.md) · [Disclaimer](https://github.com/ManagementMO/agent2learn/blob/main/DISCLAIMER.md) · [Licence](https://github.com/ManagementMO/agent2learn/blob/main/LICENSE) ·
+  [Third-party notices](https://github.com/ManagementMO/agent2learn/blob/main/THIRD_PARTY_NOTICES.md)
 
 ## Status
 
-Agent2Learn is in active development. This branch targets **0.1.2**; the pinned installers require
-that version to be available on PyPI before they can install it. A release-preparation branch is
-not evidence of publication. The repository holds the tested implementation of the full v0.1
-command surface; publication is gated on the release checks in [docs/LAUNCH.md](docs/LAUNCH.md),
-including live same-device authentication records on Windows, macOS, and Linux, and a supervised
-upload test before the submission path is enabled in any published build.
+Agent2Learn is available on [PyPI](https://pypi.org/project/agent2learn/).
+The latest [GitHub release](https://github.com/ManagementMO/agent2learn/releases/latest) carries the
+same verified wheel and source distribution. Use one install command above to start guided setup.
+If you are already installed or resuming setup, run `a2l init` in your terminal.
+
+The project remains in active development. Release checks are documented in
+[docs/LAUNCH.md](https://github.com/ManagementMO/agent2learn/blob/main/docs/LAUNCH.md), including same-device authentication and a fresh supervised upload
+test before submission capability can be enabled. **LEARN uploads remain disabled in this build.**
 
-Licensed under [Apache-2.0](LICENSE).
+Licensed under [Apache-2.0](https://github.com/ManagementMO/agent2learn/blob/main/LICENSE).

{agent2learn-0.1.2 → agent2learn-0.1.6}/README.md

@@ -7,19 +7,19 @@ a durable local vault: original files beside Markdown twins, a stable index, and
 resolve to ordinary paths and line numbers. Your agent stops guessing about your course and starts
 quoting it.
 
-> **Walkthrough recording:** not yet attached. It is a release artifact, recorded against the
-> synthetic fixture course in `tests/fixtures/` — never against a real account — so no course,
-> student, grade, or session data appears in it. Recording it is a launch gate in
-> [docs/LAUNCH.md](docs/LAUNCH.md); this README will not link a file that does not exist.
-
 ## Install
 
-Three supported options. Pick one.
+For the measured macOS workflow, repaired defects, PDF coverage gaps, and still-unverified
+human/platform checks, see the [release validation scope](https://github.com/ManagementMO/agent2learn/blob/main/docs/RELEASE_VALIDATION.md).
+Automated cross-platform tests are not a claim that every live login or course conversion works.
+
+For most students, use the command for your operating system. In a real terminal it installs
+Agent2Learn and starts guided setup immediately. If you already have uv, use the uv option below.
 
 **macOS and Linux**
 
 ```bash
-curl -fsSL https://raw.githubusercontent.com/ManagementMO/agent2learn/main/install.sh | bash
+(installer="$(curl -fsSL https://raw.githubusercontent.com/ManagementMO/agent2learn/main/install.sh)" && bash -c "$installer")
 ```
 
 **Windows (PowerShell)**
@@ -30,21 +30,35 @@ irm https://raw.githubusercontent.com/ManagementMO/agent2learn/main/install.ps1
 
 **Already have [uv](https://docs.astral.sh/uv/)?**
 
+Bash or zsh:
+
 ```bash
-uv tool install agent2learn
-a2l init
+uv tool install agent2learn && uv tool run --from agent2learn a2l init
+```
+
+Windows PowerShell:
+
+```powershell
+uv tool install agent2learn; if ($LASTEXITCODE -eq 0) { uv tool run --from agent2learn a2l init } else { throw "Agent2Learn installation failed" }
 ```
 
+These uv commands start setup through uv itself, so a newly installed `a2l` does not need to be
+on the current shell's PATH first. `uv tool install agent2learn` remains the install-only command.
+
 These commands install the published package from PyPI, not an unreleased repository checkout.
 Python 3.11–3.14 is supported. If uv selects an older system Python or your shell cannot find
-`a2l`, see [setup troubleshooting](docs/FAQ.md#setup-and-path).
+`a2l`, see [setup troubleshooting](https://github.com/ManagementMO/agent2learn/blob/main/docs/FAQ.md#setup-and-path).
 
-`install.sh` and `install.ps1` install a pinned Agent2Learn (currently 0.1.2), verify that `a2l`
-runs, and then **continue straight into interactive `a2l init`** in the same command. If they are
-run without a terminal on both ends — in CI, or through a pipe — they stop after verifying and
-print the exact next step instead of setting anything up. Neither script needs administrator
-rights, and neither creates a vault, installs agent skills, or opens a browser by itself.
-Onboarding does those things, after showing you a preview and asking.
+`install.sh` and `install.ps1` install a pinned Agent2Learn release, verify that `a2l` runs,
+and then **continue straight into interactive `a2l init`** in the same command. The
+macOS/Linux launcher waits for a successful download and preserves your terminal for setup prompts.
+If a script runs without a terminal on both ends, such as in CI or a headless agent tool, it stops
+after verification and prints the next step instead of setting anything up.
+
+Neither script needs administrator rights or creates a vault, installs agent skills, or opens a
+browser by itself. Onboarding does those things after showing you a preview and asking; you still
+complete LEARN sign-in and Duo yourself. Already-open terminals or agents may need to be restarted
+to pick up the new executable path, but you do not need to repeat completed setup.
 
 ### Agent skills
 
@@ -56,7 +70,7 @@ Separately, `npx skills add ManagementMO/agent2learn` is an optional skills-only
 skills ecosystem. It **does not install** the `a2l` engine — only the skill documents — so the
 commands they describe will not exist until you install Agent2Learn itself with one of the three
 options above. That route also runs a third-party CLI with its own network behaviour; see
-[docs/PRIVACY.md](docs/PRIVACY.md).
+[docs/PRIVACY.md](https://github.com/ManagementMO/agent2learn/blob/main/docs/PRIVACY.md).
 
 ## Then just ask
 
@@ -119,7 +133,7 @@ a2l submit COURSE ITEM FILE      preview an upload, then require your typed conf
 
 Agent2Learn is **not affiliated with, endorsed by, or supported by the University of Waterloo or
 D2L Corporation**. You are responsible for using it within your course rules and your institution's
-policies. See [DISCLAIMER.md](DISCLAIMER.md).
+policies. See [DISCLAIMER.md](https://github.com/ManagementMO/agent2learn/blob/main/DISCLAIMER.md).
 
 ## Privacy defaults
 
@@ -134,26 +148,28 @@ policies. See [DISCLAIMER.md](DISCLAIMER.md).
 
 Turning collection off does not delete what is already on disk; `a2l privacy purge` does that, and
 it previews the exact targets and requires a typed phrase first. Full detail, including every
-external request Agent2Learn can make, is in [docs/PRIVACY.md](docs/PRIVACY.md).
+external request Agent2Learn can make, is in [docs/PRIVACY.md](https://github.com/ManagementMO/agent2learn/blob/main/docs/PRIVACY.md).
 
 ## Documentation
 
-- [Install guide for agents](docs/install.md) — point your agent here and let it handle setup.
-- [Authentication](docs/AUTHENTICATION.md) — the dedicated profile, Duo, expiry, and recovery.
-- [Privacy](docs/PRIVACY.md) — what is stored, where, every network action, and how to delete it.
-- [FAQ](docs/FAQ.md) — the things that actually go wrong.
-- [Porting to another school](docs/PORTING.md) — the `School` protocol and a worked reference.
-- [What is deferred, and why](docs/FUTURE.md)
-- [Security policy](SECURITY.md) · [Disclaimer](DISCLAIMER.md) · [Licence](LICENSE) ·
-  [Third-party notices](THIRD_PARTY_NOTICES.md)
+- [Install guide for agents](https://github.com/ManagementMO/agent2learn/blob/main/docs/install.md) — point your agent here and let it handle setup.
+- [Authentication](https://github.com/ManagementMO/agent2learn/blob/main/docs/AUTHENTICATION.md) — the dedicated profile, Duo, expiry, and recovery.
+- [Privacy](https://github.com/ManagementMO/agent2learn/blob/main/docs/PRIVACY.md) — what is stored, where, every network action, and how to delete it.
+- [FAQ](https://github.com/ManagementMO/agent2learn/blob/main/docs/FAQ.md) — the things that actually go wrong.
+- [Porting to another school](https://github.com/ManagementMO/agent2learn/blob/main/docs/PORTING.md) — the `School` protocol and a worked reference.
+- [What is deferred, and why](https://github.com/ManagementMO/agent2learn/blob/main/docs/FUTURE.md)
+- [Security policy](https://github.com/ManagementMO/agent2learn/blob/main/SECURITY.md) · [Disclaimer](https://github.com/ManagementMO/agent2learn/blob/main/DISCLAIMER.md) · [Licence](https://github.com/ManagementMO/agent2learn/blob/main/LICENSE) ·
+  [Third-party notices](https://github.com/ManagementMO/agent2learn/blob/main/THIRD_PARTY_NOTICES.md)
 
 ## Status
 
-Agent2Learn is in active development. This branch targets **0.1.2**; the pinned installers require
-that version to be available on PyPI before they can install it. A release-preparation branch is
-not evidence of publication. The repository holds the tested implementation of the full v0.1
-command surface; publication is gated on the release checks in [docs/LAUNCH.md](docs/LAUNCH.md),
-including live same-device authentication records on Windows, macOS, and Linux, and a supervised
-upload test before the submission path is enabled in any published build.
+Agent2Learn is available on [PyPI](https://pypi.org/project/agent2learn/).
+The latest [GitHub release](https://github.com/ManagementMO/agent2learn/releases/latest) carries the
+same verified wheel and source distribution. Use one install command above to start guided setup.
+If you are already installed or resuming setup, run `a2l init` in your terminal.
+
+The project remains in active development. Release checks are documented in
+[docs/LAUNCH.md](https://github.com/ManagementMO/agent2learn/blob/main/docs/LAUNCH.md), including same-device authentication and a fresh supervised upload
+test before submission capability can be enabled. **LEARN uploads remain disabled in this build.**
 
-Licensed under [Apache-2.0](LICENSE).
+Licensed under [Apache-2.0](https://github.com/ManagementMO/agent2learn/blob/main/LICENSE).

{agent2learn-0.1.2 → agent2learn-0.1.6}/pyproject.toml

@@ -1,15 +1,15 @@
 [project]
 name = "agent2learn"
-version = "0.1.2"
+version = "0.1.6"
 description = "A local, source-cited course vault for AI agents"
 readme = "README.md"
 requires-python = ">=3.11"
 license = "Apache-2.0"
 license-files = ["LICENSE"]
 dependencies = [
-    "typer>=0.15,<1",
+    "typer>=0.16,<1",
     "click>=8.2,<9",
-    "rich>=13,<15",
+    "rich>=13,<16",
     "requests>=2.32,<3",
     "platformdirs>=4.2,<5",
     "keyring>=25,<27",
@@ -38,7 +38,7 @@ dev = [
     "pytest-cov>=6,<8",
     "pytest-httpserver>=1.1,<2",
     "ruff>=0.8,<1",
-    "mypy>=1.13,<2",
+    "mypy>=1.13,<3",
     "pip-audit>=2.7,<3",
     "cyclonedx-bom>=5,<8",
     "detect-secrets>=1.5,<2",
@@ -46,7 +46,7 @@ dev = [
 ]
 
 [build-system]
-requires = ["uv_build>=0.11.32,<0.12.0"]
+requires = ["uv_build>=0.11.32,<0.13.0"]
 build-backend = "uv_build"
 
 [tool.ruff]

{agent2learn-0.1.2 → agent2learn-0.1.6}/pyproject.toml.orig

@@ -1,6 +1,6 @@
 [project]
 name = "agent2learn"
-version = "0.1.2"
+version = "0.1.6"
 description = "A local, source-cited course vault for AI agents"
 readme = "README.md"
 requires-python = ">=3.11"
@@ -15,12 +15,12 @@ license-files = ["LICENSE"]
 # bump requires a release that reruns the 262-PDF acceptance corpus, explains the output
 # diff, regenerates the golden vault, and verifies bytes on all three operating systems.
 dependencies = [
-    "typer>=0.15,<1",
-    # Typer's command objects are Click groups at runtime. Keep Click explicit rather than
-    # relying on an incidental transitive dependency (the strict doctor CLI-contract test
-    # imports ``click.Group`` directly).
+    # 0.16 adds Click 8.2+ help compatibility; 0.15 permits an install that crashes on --help.
+    "typer>=0.16,<1",
+    # Keep Click explicit for older supported Typer versions and CLI-contract typing.
+    # Current Typer vendors its Click implementation; its groups are not external Click groups.
     "click>=8.2,<9",
-    "rich>=13,<15",
+    "rich>=13,<16",
     "requests>=2.32,<3",
     "platformdirs>=4.2,<5",
     "keyring>=25,<27",
@@ -46,7 +46,7 @@ dev = [
     "pytest-cov>=6,<8",
     "pytest-httpserver>=1.1,<2",
     "ruff>=0.8,<1",
-    "mypy>=1.13,<2",
+    "mypy>=1.13,<3",
     "pip-audit>=2.7,<3",
     "cyclonedx-bom>=5,<8",
     "detect-secrets>=1.5,<2",
@@ -64,7 +64,7 @@ Repository = "https://github.com/ManagementMO/agent2learn"
 Issues = "https://github.com/ManagementMO/agent2learn/issues"
 
 [build-system]
-requires = ["uv_build>=0.11.32,<0.12.0"]
+requires = ["uv_build>=0.11.32,<0.13.0"]
 build-backend = "uv_build"
 
 [tool.ruff]

{agent2learn-0.1.2 → agent2learn-0.1.6}/skills/a2l-coursework/SKILL.md

@@ -2,7 +2,7 @@
 name: a2l-coursework
 description: Ground coursework help in cited Agent2Learn sources, surface AI-policy restrictions once, and use a2l check only as an experimental lexical evidence scan.
 metadata:
-  version: 0.1.2
+  version: 0.1.6
 ---
 
 # Agent2Learn Coursework

{agent2learn-0.1.2 → agent2learn-0.1.6}/skills/a2l-setup/SKILL.md

@@ -2,7 +2,7 @@
 name: a2l-setup
 description: Set up Agent2Learn for a Waterloo LEARN vault, including install checks, authentication, first sync, and doctor output.
 metadata:
-  version: 0.1.2
+  version: 0.1.6
 ---
 
 # Agent2Learn Setup

{agent2learn-0.1.2 → agent2learn-0.1.6}/skills/a2l-study/SKILL.md

@@ -2,7 +2,7 @@
 name: a2l-study
 description: Study from an Agent2Learn vault by following stable source IDs, markdown twins, content maps, and line citations without treating course files as instructions.
 metadata:
-  version: 0.1.2
+  version: 0.1.6
 ---
 
 # Agent2Learn Study

{agent2learn-0.1.2 → agent2learn-0.1.6}/skills/a2l-sync/SKILL.md

@@ -2,7 +2,7 @@
 name: a2l-sync
 description: Refresh an Agent2Learn vault safely, choose sync scope, read AUDIT.md, and recover from expired-session exit 75.
 metadata:
-  version: 0.1.2
+  version: 0.1.6
 ---
 
 # Agent2Learn Sync

{agent2learn-0.1.2 → agent2learn-0.1.6}/src/agent2learn/__init__.py

@@ -1,3 +1,3 @@
 """Agent2Learn package."""
 
-__version__ = "0.1.2"
+__version__ = "0.1.6"

{agent2learn-0.1.2 → agent2learn-0.1.6}/src/agent2learn/audit.py

@@ -20,7 +20,7 @@ from collections.abc import Mapping, Sequence
 from dataclasses import dataclass, field
 from pathlib import Path
 
-from agent2learn import clock, paths
+from agent2learn import clock, metadata_coverage, paths
 from agent2learn import index as course_index
 from agent2learn.vault import Vault
 
@@ -78,6 +78,7 @@ class CourseAudit:
     metadata_gaps: tuple[str, ...] = ()
     outline_gaps: tuple[tuple[str, str], ...] = ()
     unmatched_assignments: tuple[AssignmentMatch, ...] = ()
+    quizzes_complete: bool = True
 
     @property
     def coverage_percent(self) -> int:
@@ -143,7 +144,10 @@ def _audit_course(vault: Vault, course_dir: Path, rows: Sequence[object]) -> Cou
         if row.get("status") == "outline_unavailable"
     )
     unmatched = _unmatched_assignments(assignments, titles)
-    metadata_gaps = tuple(gap for gap in (assignments_gap, quizzes_gap) if gap is not None)
+    quiz_coverage = metadata_coverage.read_quiz_coverage(course_dir)
+    metadata_gaps = tuple(
+        gap for gap in (assignments_gap, quizzes_gap, quiz_coverage.gap) if gap is not None
+    )
 
     first = next((row for row in rows if isinstance(row, Mapping)), {})
     return CourseAudit(
@@ -161,6 +165,7 @@ def _audit_course(vault: Vault, course_dir: Path, rows: Sequence[object]) -> Cou
         metadata_gaps=metadata_gaps,
         outline_gaps=outline_gaps,
         unmatched_assignments=unmatched,
+        quizzes_complete=quiz_coverage.status == "complete" and quizzes_gap is None,
     )
 
 
@@ -295,8 +300,8 @@ def _render(audits: Sequence[CourseAudit], stamp: str) -> str:
                 [
                     "### Metadata gaps",
                     "",
-                    "The local metadata projection could not be read completely; inventory "
-                    "counts below may be incomplete.",
+                    "Metadata coverage is incomplete or unknown; the records below may be "
+                    "cached and do not establish the current complete inventory.",
                     "",
                     *[f"- {gap}" for gap in audit.metadata_gaps],
                     "",
@@ -327,10 +332,16 @@ def _render(audits: Sequence[CourseAudit], stamp: str) -> str:
                 ["", "These are external or licensed targets. Open them in LEARN directly.", ""]
             )
 
+        quiz_inventory = (
+            f"- {_plural(audit.quizzes, 'quiz', 'quizzes')}"
+            f" ({audit.quizzes_with_due_dates} with due dates)"
+            if audit.quizzes_complete
+            else f"- Quiz inventory not confirmed; {audit.quizzes} cached quiz record(s)"
+            f" ({audit.quizzes_with_due_dates} with due dates)"
+        )
         counts = [
             f"- {_plural(audit.assignments, 'assignment')}",
-            f"- {_plural(audit.quizzes, 'quiz', 'quizzes')}"
-            f" ({audit.quizzes_with_due_dates} with due dates)",
+            quiz_inventory,
             f"- {_plural(audit.media, 'media file')}",
         ]
         lines.extend(["### Inventory", "", *counts, ""])

{agent2learn-0.1.2 → agent2learn-0.1.6}/src/agent2learn/auth/__init__.py

@@ -10,7 +10,7 @@ from urllib.parse import urljoin, urlsplit
 import requests
 
 from agent2learn import __version__, config, paths, session
-from agent2learn.errors import AuthenticationError
+from agent2learn.errors import AuthenticationError, BrowserNotFound
 from agent2learn.schools import School
 
 from . import paste
@@ -212,7 +212,10 @@ def _verified_id_from_cdp_result(value: object) -> str | None:
 
 
 def _safe_cdp_error(exc: AuthenticationError) -> str:
-    """Keep the public auth error generic while retaining a sanitized blocked hostname."""
+    """Expose fixed local prerequisite diagnostics, never arbitrary browser exception text."""
+
+    if isinstance(exc, BrowserNotFound):
+        return "Chrome or Edge was not found; install one for dedicated-browser sign-in"
 
     prefix = "authentication stopped at undeclared host "
     raw = str(exc)

{agent2learn-0.1.2 → agent2learn-0.1.6}/src/agent2learn/auth/cdp.py

@@ -21,7 +21,7 @@ import requests
 import websocket
 
 from agent2learn import config, paths
-from agent2learn.errors import AuthenticationError
+from agent2learn.errors import AuthenticationError, BrowserNotFound
 from agent2learn.schools import School
 from agent2learn.session import Session
 
@@ -493,7 +493,7 @@ def locate_browser() -> Path:
     for candidate in candidates:
         if paths.long_path(candidate).is_file():
             return candidate
-    raise AuthenticationError(
+    raise BrowserNotFound(
         "Chrome or Edge was not found; install one or use the fallback: a2l auth --paste"
     )