crewforge 0.1.0__tar.gz → 0.2.0__tar.gz

{crewforge-0.1.0/src/crewforge.egg-info → crewforge-0.2.0}/PKG-INFO

@@ -1,14 +1,14 @@
 Metadata-Version: 2.4
 Name: crewforge
-Version: 0.1.0
-Summary: One crew of LLM agents, four recipes: code review, document Q&A, API scaffolding and content — on any OpenAI-compatible provider.
+Version: 0.2.0
+Summary: One crew of LLM agents, five recipes: code review, document Q&A, API scaffolding, content and honest academic-writing help — on any OpenAI-compatible provider.
 Author: Muhammad Aiman
 License-Expression: MIT
 Project-URL: Homepage, https://github.com/aimanmalib/crewforge
 Project-URL: Repository, https://github.com/aimanmalib/crewforge
 Project-URL: Issues, https://github.com/aimanmalib/crewforge/issues
 Project-URL: Changelog, https://github.com/aimanmalib/crewforge/blob/main/CHANGELOG.md
-Keywords: llm,agents,multi-agent,openai,code-review,rag,fastapi,content-generation,cli
+Keywords: llm,agents,multi-agent,openai,code-review,rag,fastapi,content-generation,academic-writing,cli
 Classifier: Development Status :: 3 - Alpha
 Classifier: Environment :: Console
 Classifier: Intended Audience :: Developers
@@ -47,9 +47,10 @@ Dynamic: license-file
 
 # crewforge
 
-**One crew of LLM agents, four recipes — code review, document Q&A, API scaffolding and content — on any OpenAI-compatible provider, and it fails loudly instead of pretending.**
+**One crew of LLM agents, five recipes — code review, document Q&A, API scaffolding, content and honest academic-writing help — on any OpenAI-compatible provider, and it fails loudly instead of pretending.**
 
 [![CI](https://github.com/aimanmalib/crewforge/actions/workflows/ci.yml/badge.svg)](https://github.com/aimanmalib/crewforge/actions/workflows/ci.yml)
+[![PyPI](https://img.shields.io/pypi/v/crewforge)](https://pypi.org/project/crewforge/)
 ![Python 3.10–3.13](https://img.shields.io/badge/python-3.10%E2%80%933.13-blue)
 ![License: MIT](https://img.shields.io/badge/license-MIT-green)
 
@@ -58,6 +59,7 @@ crewforge review src/                                  # six reviewers + a summa
 crewforge ask handbook.pdf -q "How many leave days?" --verify
 crewforge scaffold "Todo API: list, create and delete todos" -o todo-api
 crewforge write "Why unit tests matter" -f markdown -f html
+crewforge academic feedback draft.md --brief brief.md --language ms
 ```
 
 <p align="center">
@@ -65,7 +67,7 @@ crewforge write "Why unit tests matter" -f markdown -f html
 </p>
 <p align="center"><sub>A real run of <code>crewforge review</code>: seven agents, ~25 s, provider-reported tokens. Recorded with <a href="https://github.com/charmbracelet/vhs">VHS</a> from <a href="https://github.com/aimanmalib/crewforge/blob/main/docs/demo/review.tape"><code>docs/demo/review.tape</code></a> using Gemini 3.8 Flash through an OpenAI-compatible gateway (the <code>custom</code> provider). The wait for the model is shown in full.</sub></p>
 
-## Why one crew, four recipes
+## Why one crew, many recipes
 
 crewforge started as four separate command-line tools —
 [codeguard](https://github.com/aimanmalib/codeguard),
@@ -87,6 +89,10 @@ apply to every recipe: non-streaming requests, bounded retries, usage numbers
 taken only from the provider, keys only from the environment, and no run that
 reports success when an agent produced nothing.
 
+A fifth recipe, `academic`, replaces a standalone "academic writing pipeline"
+that drafted essays and marketed itself as AI-detector-resistant. Only the
+honest parts were kept; see [docs/porting.md](https://github.com/aimanmalib/crewforge/blob/main/docs/porting.md#academic-from-hermes-mimo-writer).
+
 ## 60-second quickstart
 
 ```bash
@@ -128,6 +134,10 @@ tables are trimmed to their totals.
 | `scaffold "DESCRIPTION"` | parser → schema → router, model, auth, test_gen, doc_gen (parallel) → validator | A runnable FastAPI project with `openapi.json`, tests and docs |
 | `validate PATH` | validator (no LLM) | Spec check; for a project directory, also checks the app serves every spec operation |
 | `write "TOPIC"` | research → outline → writer → SEO → editor ⇄ quality → translator → publisher | Article as Markdown/HTML/JSON, social snippets, a usage file |
+| `academic plan BRIEF` | brief → planner | Research questions, an outline with a word budget that adds up, sources to find, suggested readings labelled UNVERIFIED |
+| `academic feedback DRAFT` | structure, argument, clarity, citations (parallel) + a citation cross-check (no LLM) | Comments tied to quoted text in your draft; no rewriting |
+| `academic cite REFS` | formatter | Your references in APA 7, IEEE or Harvard; nothing added, missing details listed |
+| `academic disclose` | none (no LLM) | An AI-use statement built from the record of what the tool did |
 
 ### `review`
 
@@ -212,6 +222,68 @@ length (`--words` is a target given to the agents, not a hard limit), and the
 research agent has no web access, so statistics it quotes come from the model's
 training data. Review articles before publishing them.
 
+### `academic`
+
+An honest assistant for a student's or researcher's **own** work. It plans,
+comments and formats; it does not write the assignment, does not invent
+references, and has no "humanise" or AI-detector-evasion feature. Output in
+English or Malaysian Malay (`--language en|ms`; Malay output is checked for
+Indonesian forms such as *karena* or *kualitas*). Every step is recorded
+(file names and hashes, never content or keys) so `disclose` can write the
+AI-use statement many universities now ask for.
+
+```text
+$ crewforge academic plan brief.md -f markdown          # a 1,200-word TSL3013 essay brief
+# Assignment plan: Regulated Integration over Prohibition: ...
+- **Target length:** 1,200 words
+- **Citation style:** APA 7th edition
+## Research questions
+1. What pedagogical benefits does generative AI offer secondary school students completing homework tasks?
+...
+| # | Section | Words | Purpose |
+| 1 | Introduction | 150 | ... |   ...   Total: 1,200 words
+## Suggested readings (UNVERIFIED)
+> The AI proposed these. They may not exist, or may not say what is claimed. ...
+- ⚠ **UNVERIFIED: proposed or altered by the AI, check it** — Holmes, W., & Tuomi, I. (2022). ...
+total   2 calls   9,034 tokens (provider-reported)
+
+$ crewforge academic feedback draft.md --brief brief.md -f markdown
+## Argument
+- **Paragraph 2** (major): “Everyone knows that tuition in Malaysia is very expensive.”
+  - This is an unsupported generalization and an informal assertion rather than an evidence-based academic claim.
+  - *Suggestion:* Replace the appeal to common knowledge with scholarly evidence or empirical data ...
+## Citation check (no AI)
+**Cited in the text but not in the reference list:**
+- Lim, 2024
+total   4 calls   17,892 tokens (provider-reported)
+
+$ crewforge academic cite refs.txt --style ieee --language ms -f markdown
+[4] W. Holmes, M. Bialik, and C. Fadel, *Artificial Intelligence in Education: ...*. Center for Curriculum Redesign, 2019, ISBN: 978-1-7941-1293-0.
+  - ⚠ nombor semak ISBN salah: 9781794112930
+  - Butiran yang perlu ditambah: tempat penerbitan
+
+$ crewforge academic disclose --language ms             # no LLM, no key
+## Pernyataan Penggunaan AI
+Saya menggunakan crewforge academic (versi 0.1.0), sebuah alat berbantukan AI, dengan model gemini-3.8-flash (melalui gateway serasi OpenAI), bagi bahagian tugasan berikut:
+- 22 Oktober 2026: Perancangan: alat ini menukar taklimat tugasan (brief-ms.md) kepada 3 soalan kajian, rangka 5 bahagian dan peruntukan perkataan. ...
+```
+
+How sources are labelled — there is no "verified" status, because the tool
+cannot check that a source exists:
+
+| Source | Label |
+|---|---|
+| You supplied it | *supplied by you (not checked online)* |
+| The model proposed it, with a DOI, URL or valid ISBN | **UNVERIFIED** — check it |
+| The model proposed it with none of those | **LIKELY FABRICATED** |
+| The model added a DOI/URL/ISBN to a reference you supplied | **UNVERIFIED**, naming the added identifier |
+
+A DOI the model proposes can be made up too, which is why even those stay
+unverified. All text the student sees lives in
+`src/crewforge/recipes/academic/locales/{en,ms}.yaml`; defaults (language,
+style, log file, input size, suggestion length) are the `academic:` section of
+`crewforge.yaml`.
+
 ## Providers
 
 The table is checked against the code by `tests/test_readme.py`.
@@ -239,13 +311,14 @@ An unknown provider name is a configuration error, not a silent fallback.
 
 ```mermaid
 flowchart LR
-    CLI["crewforge CLI<br/>review · ask · summarize · ingest<br/>scaffold · validate · write"] --> R
+    CLI["crewforge CLI<br/>review · ask · summarize · ingest<br/>scaffold · validate · write<br/>academic plan · feedback · cite · disclose"] --> R
 
     subgraph R["Recipes (src/crewforge/recipes)"]
         RV[review]
         DC[docs]
         AP[api]
         CT[content]
+        AC[academic]
     end
 
     R --> CREW
@@ -317,7 +390,7 @@ Validation: INVALID · 0/100 (0 operation(s) checked, 1 error(s), 0 warning(s))
   ERROR: No endpoints defined (paths is empty)
 ```
 
-| Situation | Before (the four tools) | crewforge |
+| Situation | Before (the earlier tools) | crewforge |
 |---|---|---|
 | Every reviewer returns nothing | "Review Complete!", exit 0 | exit 1, names each reviewer |
 | A reviewer answers in prose | treated as a review | exit 1, `agent 'style' (parse)` |
@@ -327,6 +400,9 @@ Validation: INVALID · 0/100 (0 operation(s) checked, 1 error(s), 0 warning(s))
 | Quality never reaches the threshold | article published | exit 1 unless `--accept-below-threshold` |
 | Question matches no passage | model asked to answer from nothing | exit 1, names `retriever`, no LLM call |
 | Generated code does not compile | template silently reused | exit 1, names the agent |
+| Model proposes a reference with no DOI/URL/ISBN | hermes-mimo-writer put invented "plausible" references in the bibliography | labelled LIKELY FABRICATED, never verified |
+| Formatter skips a reference you supplied | n/a | exit 1, names `formatter` |
+| `disclose` finds no recorded AI use | n/a | exit 1, no statement written |
 
 Exit codes: `0` success · `1` run failed (agent, validator or provider) · `2`
 usage or configuration error (bad flag, unknown provider, missing key).

{crewforge-0.1.0 → crewforge-0.2.0}/README.md

@@ -1,8 +1,9 @@
 # crewforge
 
-**One crew of LLM agents, four recipes — code review, document Q&A, API scaffolding and content — on any OpenAI-compatible provider, and it fails loudly instead of pretending.**
+**One crew of LLM agents, five recipes — code review, document Q&A, API scaffolding, content and honest academic-writing help — on any OpenAI-compatible provider, and it fails loudly instead of pretending.**
 
 [![CI](https://github.com/aimanmalib/crewforge/actions/workflows/ci.yml/badge.svg)](https://github.com/aimanmalib/crewforge/actions/workflows/ci.yml)
+[![PyPI](https://img.shields.io/pypi/v/crewforge)](https://pypi.org/project/crewforge/)
 ![Python 3.10–3.13](https://img.shields.io/badge/python-3.10%E2%80%933.13-blue)
 ![License: MIT](https://img.shields.io/badge/license-MIT-green)
 
@@ -11,6 +12,7 @@ crewforge review src/                                  # six reviewers + a summa
 crewforge ask handbook.pdf -q "How many leave days?" --verify
 crewforge scaffold "Todo API: list, create and delete todos" -o todo-api
 crewforge write "Why unit tests matter" -f markdown -f html
+crewforge academic feedback draft.md --brief brief.md --language ms
 ```
 
 <p align="center">
@@ -18,7 +20,7 @@ crewforge write "Why unit tests matter" -f markdown -f html
 </p>
 <p align="center"><sub>A real run of <code>crewforge review</code>: seven agents, ~25 s, provider-reported tokens. Recorded with <a href="https://github.com/charmbracelet/vhs">VHS</a> from <a href="https://github.com/aimanmalib/crewforge/blob/main/docs/demo/review.tape"><code>docs/demo/review.tape</code></a> using Gemini 3.8 Flash through an OpenAI-compatible gateway (the <code>custom</code> provider). The wait for the model is shown in full.</sub></p>
 
-## Why one crew, four recipes
+## Why one crew, many recipes
 
 crewforge started as four separate command-line tools —
 [codeguard](https://github.com/aimanmalib/codeguard),
@@ -40,6 +42,10 @@ apply to every recipe: non-streaming requests, bounded retries, usage numbers
 taken only from the provider, keys only from the environment, and no run that
 reports success when an agent produced nothing.
 
+A fifth recipe, `academic`, replaces a standalone "academic writing pipeline"
+that drafted essays and marketed itself as AI-detector-resistant. Only the
+honest parts were kept; see [docs/porting.md](https://github.com/aimanmalib/crewforge/blob/main/docs/porting.md#academic-from-hermes-mimo-writer).
+
 ## 60-second quickstart
 
 ```bash
@@ -81,6 +87,10 @@ tables are trimmed to their totals.
 | `scaffold "DESCRIPTION"` | parser → schema → router, model, auth, test_gen, doc_gen (parallel) → validator | A runnable FastAPI project with `openapi.json`, tests and docs |
 | `validate PATH` | validator (no LLM) | Spec check; for a project directory, also checks the app serves every spec operation |
 | `write "TOPIC"` | research → outline → writer → SEO → editor ⇄ quality → translator → publisher | Article as Markdown/HTML/JSON, social snippets, a usage file |
+| `academic plan BRIEF` | brief → planner | Research questions, an outline with a word budget that adds up, sources to find, suggested readings labelled UNVERIFIED |
+| `academic feedback DRAFT` | structure, argument, clarity, citations (parallel) + a citation cross-check (no LLM) | Comments tied to quoted text in your draft; no rewriting |
+| `academic cite REFS` | formatter | Your references in APA 7, IEEE or Harvard; nothing added, missing details listed |
+| `academic disclose` | none (no LLM) | An AI-use statement built from the record of what the tool did |
 
 ### `review`
 
@@ -165,6 +175,68 @@ length (`--words` is a target given to the agents, not a hard limit), and the
 research agent has no web access, so statistics it quotes come from the model's
 training data. Review articles before publishing them.
 
+### `academic`
+
+An honest assistant for a student's or researcher's **own** work. It plans,
+comments and formats; it does not write the assignment, does not invent
+references, and has no "humanise" or AI-detector-evasion feature. Output in
+English or Malaysian Malay (`--language en|ms`; Malay output is checked for
+Indonesian forms such as *karena* or *kualitas*). Every step is recorded
+(file names and hashes, never content or keys) so `disclose` can write the
+AI-use statement many universities now ask for.
+
+```text
+$ crewforge academic plan brief.md -f markdown          # a 1,200-word TSL3013 essay brief
+# Assignment plan: Regulated Integration over Prohibition: ...
+- **Target length:** 1,200 words
+- **Citation style:** APA 7th edition
+## Research questions
+1. What pedagogical benefits does generative AI offer secondary school students completing homework tasks?
+...
+| # | Section | Words | Purpose |
+| 1 | Introduction | 150 | ... |   ...   Total: 1,200 words
+## Suggested readings (UNVERIFIED)
+> The AI proposed these. They may not exist, or may not say what is claimed. ...
+- ⚠ **UNVERIFIED: proposed or altered by the AI, check it** — Holmes, W., & Tuomi, I. (2022). ...
+total   2 calls   9,034 tokens (provider-reported)
+
+$ crewforge academic feedback draft.md --brief brief.md -f markdown
+## Argument
+- **Paragraph 2** (major): “Everyone knows that tuition in Malaysia is very expensive.”
+  - This is an unsupported generalization and an informal assertion rather than an evidence-based academic claim.
+  - *Suggestion:* Replace the appeal to common knowledge with scholarly evidence or empirical data ...
+## Citation check (no AI)
+**Cited in the text but not in the reference list:**
+- Lim, 2024
+total   4 calls   17,892 tokens (provider-reported)
+
+$ crewforge academic cite refs.txt --style ieee --language ms -f markdown
+[4] W. Holmes, M. Bialik, and C. Fadel, *Artificial Intelligence in Education: ...*. Center for Curriculum Redesign, 2019, ISBN: 978-1-7941-1293-0.
+  - ⚠ nombor semak ISBN salah: 9781794112930
+  - Butiran yang perlu ditambah: tempat penerbitan
+
+$ crewforge academic disclose --language ms             # no LLM, no key
+## Pernyataan Penggunaan AI
+Saya menggunakan crewforge academic (versi 0.1.0), sebuah alat berbantukan AI, dengan model gemini-3.8-flash (melalui gateway serasi OpenAI), bagi bahagian tugasan berikut:
+- 22 Oktober 2026: Perancangan: alat ini menukar taklimat tugasan (brief-ms.md) kepada 3 soalan kajian, rangka 5 bahagian dan peruntukan perkataan. ...
+```
+
+How sources are labelled — there is no "verified" status, because the tool
+cannot check that a source exists:
+
+| Source | Label |
+|---|---|
+| You supplied it | *supplied by you (not checked online)* |
+| The model proposed it, with a DOI, URL or valid ISBN | **UNVERIFIED** — check it |
+| The model proposed it with none of those | **LIKELY FABRICATED** |
+| The model added a DOI/URL/ISBN to a reference you supplied | **UNVERIFIED**, naming the added identifier |
+
+A DOI the model proposes can be made up too, which is why even those stay
+unverified. All text the student sees lives in
+`src/crewforge/recipes/academic/locales/{en,ms}.yaml`; defaults (language,
+style, log file, input size, suggestion length) are the `academic:` section of
+`crewforge.yaml`.
+
 ## Providers
 
 The table is checked against the code by `tests/test_readme.py`.
@@ -192,13 +264,14 @@ An unknown provider name is a configuration error, not a silent fallback.
 
 ```mermaid
 flowchart LR
-    CLI["crewforge CLI<br/>review · ask · summarize · ingest<br/>scaffold · validate · write"] --> R
+    CLI["crewforge CLI<br/>review · ask · summarize · ingest<br/>scaffold · validate · write<br/>academic plan · feedback · cite · disclose"] --> R
 
     subgraph R["Recipes (src/crewforge/recipes)"]
         RV[review]
         DC[docs]
         AP[api]
         CT[content]
+        AC[academic]
     end
 
     R --> CREW
@@ -270,7 +343,7 @@ Validation: INVALID · 0/100 (0 operation(s) checked, 1 error(s), 0 warning(s))
   ERROR: No endpoints defined (paths is empty)
 ```
 
-| Situation | Before (the four tools) | crewforge |
+| Situation | Before (the earlier tools) | crewforge |
 |---|---|---|
 | Every reviewer returns nothing | "Review Complete!", exit 0 | exit 1, names each reviewer |
 | A reviewer answers in prose | treated as a review | exit 1, `agent 'style' (parse)` |
@@ -280,6 +353,9 @@ Validation: INVALID · 0/100 (0 operation(s) checked, 1 error(s), 0 warning(s))
 | Quality never reaches the threshold | article published | exit 1 unless `--accept-below-threshold` |
 | Question matches no passage | model asked to answer from nothing | exit 1, names `retriever`, no LLM call |
 | Generated code does not compile | template silently reused | exit 1, names the agent |
+| Model proposes a reference with no DOI/URL/ISBN | hermes-mimo-writer put invented "plausible" references in the bibliography | labelled LIKELY FABRICATED, never verified |
+| Formatter skips a reference you supplied | n/a | exit 1, names `formatter` |
+| `disclose` finds no recorded AI use | n/a | exit 1, no statement written |
 
 Exit codes: `0` success · `1` run failed (agent, validator or provider) · `2`
 usage or configuration error (bad flag, unknown provider, missing key).

{crewforge-0.1.0 → crewforge-0.2.0}/pyproject.toml

@@ -4,14 +4,14 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "crewforge"
-version = "0.1.0"
-description = "One crew of LLM agents, four recipes: code review, document Q&A, API scaffolding and content — on any OpenAI-compatible provider."
+version = "0.2.0"
+description = "One crew of LLM agents, five recipes: code review, document Q&A, API scaffolding, content and honest academic-writing help — on any OpenAI-compatible provider."
 readme = "README.md"
 license = "MIT"
 license-files = ["LICENSE"]
 requires-python = ">=3.10"
 authors = [{ name = "Muhammad Aiman" }]
-keywords = ["llm", "agents", "multi-agent", "openai", "code-review", "rag", "fastapi", "content-generation", "cli"]
+keywords = ["llm", "agents", "multi-agent", "openai", "code-review", "rag", "fastapi", "content-generation", "academic-writing", "cli"]
 classifiers = [
     "Development Status :: 3 - Alpha",
     "Environment :: Console",
@@ -61,7 +61,7 @@ crewforge = "crewforge.cli:main"
 where = ["src"]
 
 [tool.setuptools.package-data]
-crewforge = ["py.typed", "recipes/api/templates/*.j2"]
+crewforge = ["py.typed", "recipes/api/templates/*.j2", "recipes/academic/locales/*.yaml"]
 
 [tool.pytest.ini_options]
 testpaths = ["tests"]

crewforge-0.2.0/src/crewforge/__init__.py

@@ -0,0 +1,3 @@
+"""crewforge — one crew of LLM agents, five recipes, any OpenAI-compatible provider."""
+
+__version__ = "0.2.0"

{crewforge-0.1.0 → crewforge-0.2.0}/src/crewforge/cli.py

@@ -1,4 +1,4 @@
-"""``crewforge`` command line: global provider options + the four recipes."""
+"""``crewforge`` command line: global provider options + the recipes."""
 
 from __future__ import annotations
 
@@ -13,6 +13,7 @@ from .cli_support import EXIT_USAGE, console, err_console, fail
 from .core.config import LLMConfig, Settings
 from .core.errors import ConfigError
 from .core.providers import ENV_API_KEY, PROVIDERS
+from .recipes.academic.cli import commands as academic_commands
 from .recipes.api.cli import commands as api_commands
 from .recipes.content.cli import commands as content_commands
 from .recipes.docs.cli import commands as docs_commands
@@ -41,7 +42,7 @@ from .recipes.review.cli import commands as review_commands
 @click.option("--verbose", "-v", is_flag=True, help="Log HTTP retries and agent progress.")
 @click.pass_context
 def main(ctx, provider, model, base_url, config_path, verbose):
-    """One crew of LLM agents, four recipes, any OpenAI-compatible provider.
+    """One crew of LLM agents, five recipes, any OpenAI-compatible provider.
 
     \b
     Recipes:
@@ -52,6 +53,7 @@ def main(ctx, provider, model, base_url, config_path, verbose):
       scaffold   natural language -> FastAPI project + OpenAPI spec
       validate   check an OpenAPI spec (and a scaffolded app) offline
       write      research -> outline -> write -> SEO -> edit -> quality -> publish
+      academic   plan / feedback / cite / disclose for your own assignment
 
     API keys are read from the environment only (e.g. OPENAI_API_KEY) and are
     never written to disk. Exit codes: 0 ok, 1 run failed, 2 usage/config error.
@@ -120,7 +122,13 @@ def init(ctx, output, force):
     )
 
 
-for _cmd in (*review_commands, *docs_commands, *api_commands, *content_commands):
+for _cmd in (
+    *review_commands,
+    *docs_commands,
+    *api_commands,
+    *content_commands,
+    *academic_commands,
+):
     main.add_command(_cmd)
 
 

{crewforge-0.1.0 → crewforge-0.2.0}/src/crewforge/core/config.py

@@ -155,6 +155,7 @@ class Settings(BaseModel):
     docs: dict[str, Any] = Field(default_factory=dict)
     api: dict[str, Any] = Field(default_factory=dict)
     content: dict[str, Any] = Field(default_factory=dict)
+    academic: dict[str, Any] = Field(default_factory=dict)
 
     @classmethod
     def from_yaml(cls, path: str | Path) -> Settings:

crewforge-0.2.0/src/crewforge/recipes/__init__.py

@@ -0,0 +1 @@
+"""The recipes: review (code), docs (Q&A), api (scaffolding), content (articles), academic."""

crewforge-0.2.0/src/crewforge/recipes/academic/__init__.py

@@ -0,0 +1,46 @@
+"""``academic`` recipe: an honest academic writing assistant.
+
+It helps a student or researcher plan and improve their OWN work: parse a
+brief into a plan, comment on a draft, format references the user supplies,
+and write an AI-use disclosure from a record of what the tool actually did.
+It does not write assignments, does not invent references, and has no
+"humanise" or detector-evasion feature (see docs/porting.md).
+"""
+
+from .agents import (
+    ArgumentAgent,
+    BriefAgent,
+    CitationReviewAgent,
+    ClarityAgent,
+    FormatterAgent,
+    PlannerAgent,
+    StructureAgent,
+)
+from .citations import SourceCheck, analyse_draft, check_formatted, check_source, identifiers
+from .disclosure import build_statement, make_record, read_records
+from .pipeline import AcademicCrew, CiteResult, FeedbackResult, PlanResult, allocate_words
+from .settings import AcademicSettings
+
+__all__ = [
+    "AcademicCrew",
+    "AcademicSettings",
+    "ArgumentAgent",
+    "BriefAgent",
+    "CitationReviewAgent",
+    "CiteResult",
+    "ClarityAgent",
+    "FeedbackResult",
+    "FormatterAgent",
+    "PlanResult",
+    "PlannerAgent",
+    "SourceCheck",
+    "StructureAgent",
+    "allocate_words",
+    "analyse_draft",
+    "build_statement",
+    "check_formatted",
+    "check_source",
+    "identifiers",
+    "make_record",
+    "read_records",
+]