lit-review-orchestrator

name

lit-review-orchestrator

description

Master controller for the lit-review pipeline, driven by a document. Give it a .tex or .docx file describing an article — a full manuscript, an abstract, or a proposal — and it extracts a search plan, runs Undermind (an automated Playwright driver in Classic mode) and Google Scholar (SearchAPI.io), then merges, deduplicates, and screens the results. Only use this skill when explicitly requested — e.g., the user says "run lit-review-orchestrator", "lit-review-orchestrator", or "/lit-review-orchestrator". Do NOT auto-trigger on general literature review requests.

Lit-Review Orchestrator

Run the literature-review pipeline from a single command, starting from a document that describes your article.

Input: a .tex or .docx document — a full manuscript, an abstract, or any text describing the article's content. Output: a deduplicated, relevance-screened master list (JSON + RIS), plus the extracted search plan and all intermediate stage files.

Quick Start

The orchestrator.py commands below are the autonomous fallback (reasoning on the Sonnet/DeepSeek API). When an agent runs this skill interactively, use the agent-driven flow instead; see How it runs below. That flow performs non-browser reasoning at the agent layer with no Anthropic API key, using the platform routing in docs/claude-code.md or docs/codex.md.

pip install -r requirements.txt                                  # one-time
cp lit-review-pipeline.env.example ~/.lit-review-pipeline.env    # then fill in keys

# From a full manuscript
python scripts/orchestrator.py paper.docx --output-dir ~/lit-reviews/mypaper

# From just an abstract (any .tex/.docx describing the article works)
python scripts/orchestrator.py abstract.tex --output-dir ~/lit-reviews/mypaper

# Add opt-in sources; DOI-only dedup
python scripts/orchestrator.py paper.tex --ssrn --nber --no-llm --output-dir out

# Escape hatch: run from a raw query string (skips Stage 0 extraction)
python scripts/orchestrator.py --query "dual-class shares cost of equity" --output-dir out

Pipeline

Stage	Name	Default	What it does
0	Extract	on	Parse the document; the agent derives the research question, an Undermind brief, a Scholar Labs question, and Google Scholar queries
1	Undermind	on	Automated Undermind.ai Classic deep search from the brief (Playwright; signs in with stored credentials)
2	Scholar Labs	opt-in	Google Scholar Labs deep search via --scholarlabs (Playwright; stored Google login). Off by default — Google rate-limits its Cite/BibTeX export under automation, so it often defers
2b	Deep Research	on	Gemini Deep Research Agent (Interactions API; GEMINI_API_KEY) — alternative API-driven deep search
4a	Google Scholar	on	SearchAPI.io Google Scholar, driven by the extracted queries
4b	Supplementary	off	SSRN / NBER / HeinOnline / forthcoming (--ssrn --nber --heinonline --forthcoming)
4c	Citation chain	off	Semantic Scholar (--citation-chain; needs DOI-bearing seeds)
5	Dedup	on	Merge all outputs; metadata enrichment + DOI and LLM fuzzy dedup
5b	Verify	on	Cross-check every paper against OpenAlex / Crossref / Semantic Scholar and drop any none can confirm (anti-hallucination); if an index outage leaves >30% of papers uncheckable, it keeps everything and warns instead of dropping. --no-verify keeps all; dropped papers saved to stage5_merged_unverified.json
6	Screen	on	Abstract relevance screening against the research question

Stage 0 runs first; Stages 1, 2b, and 4a run concurrently (Stage 2 Scholar Labs joins them only with --scholarlabs, and 4b when opted in); 4c follows them; then 5, then 6.

How it runs: agent-driven (default) vs autonomous fallback

LLM work follows one routing rule: the agent-driven flow uses no Anthropic API for the pipeline's reasoning steps. Stage 0 extraction, Stage 4 query condensation, Stage 5 dedup judgments, Stage 6 screening, keyless web search, and Undermind clarifying answers run at the agent layer. The exact model and delegation policy depends on the host agent:

• Claude Code: read docs/claude-code.md.
• Codex: read docs/codex.md.

Undermind is a subprocess that owns the live browser, so its clarifying questions come back to the agent through a small file handshake (--answers-dir): the driver writes each question to a file and types whatever answer you drop back. The Sonnet/DeepSeek API path remains the autonomous fallback (orchestrator.py) for unattended runs with no agent present.

Because a Python subprocess cannot spawn subagents, each reasoning script exposes an emit/ingest seam: the script does the deterministic work (parsing, candidate-pair generation, validation, enrichment, merge, all file output) and hands only the LLM step out to you in the middle. Each script also keeps its in-script Sonnet/DeepSeek API path as an autonomous fallback for unattended runs, so the same files support both interactive and unattended use.

Platform routing (agent-driven flow)

The shared pipeline uses named reasoning roles. Map those roles to the host platform before dispatching subagents or doing web-search work.

Role / stage	Strong-reasoning route	Cost-conscious route
Orchestrator: coordinate the run, parse the GUI config, fan out, merge	Host platform's strongest interactive model	Parent session default
Stage 0: extract the search plan	Strong-reasoning route	Avoid downgrading unless the user requests a fast pass
Stage 4d: keyless web-search fan-out	Strong-reasoning route	Avoid downgrading; recall and precision matter
Stage 6: relevance re-ranking	Strong-reasoning route	Avoid downgrading; this determines final ranking
Stage 4a query writing, Stage 5 dedup judgments, Undermind clarifying answers	Strong-reasoning route when accuracy is prioritized	Cheaper platform worker model; escalate uncertain cases

For Claude Code, the strong route is Opus 4.8 and the cheaper worker route is Sonnet 4.6. For Codex, the strong keyless route is gpt-5.5 with xhigh reasoning, and lower-stakes batch judgments use cheaper Codex workers such as gpt-5.4-mini at medium or high reasoning. See docs/claude-code.md and docs/codex.md for the complete mapping.

The GUI still emits the legacy field "all_opus". Interpret "all_opus": true as the high-accuracy platform profile: Claude Code uses Opus for all delegated reasoning, while Codex uses gpt-5.5 with xhigh reasoning for all delegated reasoning. When it is false or absent, follow the platform's default split.

Interactive entry (GUI)

When the skill is triggered interactively, open the settings dialog first, let the user choose the input and options, then run the agent-driven stages below honouring what it returns:

python scripts/lit_review_gui.py --config-out OUT/gui_config.json   # blocks until Run/Cancel

The window has a Browse field for the document (or a raw-query box), an output folder, the search channels as checkboxes — keyed (Undermind / Deep Research / Google Scholar checked; Scholar Labs unchecked — opt-in) and keyless (Free index search / Web search, both on) — supplementary sources (SSRN checked by default, NBER, HeinOnline) plus citation chaining, Processing (Deduplicate / Verify sources / Screen / DOI-only), and an Advanced group (Quick mode, Max chars, and a legacy Use Opus for all tasks toggle that is off by default. Leave it off to run the platform's default routing split; check it to request the high-accuracy route for every delegated reasoning stage. On Run it writes the settings to --config-out and echoes them to stdout between ===LITREVIEW_CONFIG_BEGIN=== and ===LITREVIEW_CONFIG_END=== (exit 0); Cancel or closing the window exits 2 — abort the run. Parse that JSON and map it onto the stages: skip a channel set false, set output_dir, run Scholar Labs / supplementary / citation and the keyless freesearch (Stage 4e) / websearch (Stage 4d) channels when true, pass --no-verify to dedup when verify is false and --no-llm when no_llm is true, skip dedup/screen when false, pass max_chars to extraction, and, when all_opus is true, use the high-accuracy platform profile instead of the default split (see Platform routing). The GUI runs nothing itself and calls no API. Shape:

{"document":"…","query":"","output_dir":"…",
 "channels":{"undermind":true,"deepresearch":true,"scholar":true,"scholarlabs":false,"freesearch":true,"websearch":true},
 "supplementary":{"ssrn":true,"nber":false,"heinonline":false},
 "citation_chain":false,"top_seeds":20,
 "dedup":true,"verify":true,"screen":true,"no_llm":false,"quick":false,"max_chars":30000,"all_opus":false}

Agent-driven run (the default; you orchestrate)

Pick an output dir OUT. Run the deterministic stages as subprocesses and do the reasoning stages with the host platform routing described above. In Codex, spawn subagents only after explicit user authorization for parallel agent work. Substitute <doc> and the extracted <research_question>.

Stage 0: extract (strong-reasoning route):

python scripts/extract_search_plan.py <doc> --emit-prompt OUT/extract_prompt.txt -o OUT/search_plan.json
# Read OUT/extract_prompt.txt, produce the plan JSON with the platform's strong-reasoning route, write OUT/plan.json.
python scripts/extract_search_plan.py <doc> --plan-file OUT/plan.json -o OUT/search_plan.json

The plan JSON must carry extract_search_plan.py's REQUIRED_KEYS; --plan-file validates them (exit 1 on a bad plan) and writes search_plan.json/.md, scholar_queries.json, undermind_brief.txt, scholarlabs_query.txt.

Stages 1 / 2b / 4a — search (subprocesses; run concurrently, background + Monitor). Stage 2 Scholar Labs is opt-in — run it only on request (see below):

python undermind-search/scripts/undermind_search.py --brief-file OUT/undermind_brief.txt \
    -o OUT/stage1_undermind.json --debug-dir OUT/debug_undermind --answers-dir OUT/undermind_clarify
# Agent-in-the-loop, no API: while it runs, watch OUT/undermind_clarify for clarify_request_<n>.json.
# When one appears, answer the question with the platform worker route (strong route when all_opus is true), grounded in the
# brief and the Stage-0 undermind_clarifications, and write OUT/undermind_clarify/clarify_answer_<n>.json = {"answer": "..."}.
# Practical pattern: launch a background job that blocks until the request file exists (so you are
# notified), answer it, then re-arm for the next turn. Undermind is interactive in this mode.
# Opt-in only (Scholar Labs): Google rate-limits its Cite export under automation, so skip it by
# default and run this line only when asked / retrying from a fresh session:
python scholarlabs-search/scripts/scholarlabs_search.py --query-file OUT/scholarlabs_query.txt \
    --research-question "<research_question>" -o OUT/stage2_scholarlabs.json --hidden --debug-dir OUT/debug_scholarlabs
python deepresearch-search/scripts/deepresearch_search.py --query-file OUT/undermind_brief.txt \
    --research-question "<research_question>" -o OUT/stage2b_deepresearch.json --debug-dir OUT/debug_deepresearch  # Gemini Deep Research (GEMINI_API_KEY), pure subprocess
python supplementary-search/scripts/supplementary_search.py --scholar \
    --queries-file OUT/scholar_queries.json -o OUT/stage4a_scholar.json --debug-dir OUT/debug_scholar

Passing --queries-file (the agent-written queries, using the platform worker route by default and the strong route when all_opus is true) bypasses the in-script condense_query fallback in supplementary-search. For a raw-query agent run (no document, hence no Stage 0 to produce scholar_queries.json), first have the agent write that file (a short JSON array of query strings) and pass it the same way, or add --no-condense. Either route keeps the agent path free of the in-script API call.

Web search (keyless agent-driven channel, and a useful add-on alongside the keyed channels). Subagent fan-out, so the raw web text stays out of your context. This is the keyless search route in Platform routing. Emit a batched task plan, dispatch one strong-reasoning subagent per batch when parallel agents are authorized, then merge:

python websearch-search/scripts/websearch_ingest.py --emit-tasks \
    --queries-file OUT/scholar_queries.json --research-question "<research_question>" \
    --batch-size 3 -o OUT/websearch_tasks.json
# Dispatch one strong-reasoning worker per tasks[k]: hand it the system_prompt + its queries;
# each uses the platform's web-search/fetch tools over its queries and writes
# OUT/websearch_results_batch_<id>.json (only title required; never invent fields; do
# not fetch scholar.google.com). Then merge the partials:
python websearch-search/scripts/websearch_ingest.py \
    --results OUT/websearch_results_batch_*.json -o OUT/stage4d_websearch.json

This writes stage4d_websearch.json (source="websearch"), deduped by title with best-effort keyless Crossref DOI fill, which the dedup --inputs glob below picks up. The hits are real web results, so keep Stage 5b verification ON. For a few queries you can skip the fan-out and ingest a single websearch_results.json. Empty input defers (WEBSEARCH_DEFERRED). Full recipe: websearch-search/SKILL.md.

Free index search (keyless; pairs with web search for the no-key fallback). A plain keyless subprocess that searches OpenAlex / Crossref / Semantic Scholar with the Stage-0 queries:

python freesearch-search/scripts/freesearch_search.py \
    --queries-file OUT/scholar_queries.json -o OUT/stage4e_freesearch.json

This writes stage4e_freesearch.json (real index records, source set per index), which the dedup --inputs glob below also picks up. No key needed; see freesearch-search/SKILL.md.

Stage 5: dedup (platform worker route by default; strong route when all_opus):

python lit-dedup/scripts/lit_dedup.py --inputs OUT/stage[0-9]*.json --emit-pairs OUT/dedup_pairs.json -o OUT/stage5_merged.json
# Read OUT/dedup_pairs.json; for each pairs[k] = {i, j, a, b} decide if a and b are the
# same paper. Fan out across parallel platform workers for large pair sets when authorized. Write
# OUT/dedup_verdicts.json = [{"i":N,"j":N,"decision":"yes|no","confidence":"high|medium|low","rationale":"..."}].
python lit-dedup/scripts/lit_dedup.py --ingest-verdicts OUT/dedup_pairs.json OUT/dedup_verdicts.json -o OUT/stage5_merged.json

Exclude stage5_* / stage6_* from the --inputs glob. If dedup_pairs.json has no pairs, write [] to the verdicts file and still run --ingest-verdicts.

Stage 6: screen (the re-ranker; strong-reasoning route):

python lit-screen/scripts/lit_screen.py --input OUT/stage5_merged.json --query "<research_question>" \
    --emit-tasks OUT/screen_tasks.json -o OUT/stage6_screened.json
# Read OUT/screen_tasks.json = {system_prompt, query, tasks:[{index, user_message}]}.
# Score each task following system_prompt. Fan out across parallel strong-reasoning workers in batches when authorized. Write
# OUT/screen_results.json = [{"index":N,"relevance_score":1-10,"rationale":"..","paper_type":"..","identification_strategy":"..","relationship":".."}].
python lit-screen/scripts/lit_screen.py --input OUT/stage5_merged.json --ingest-results OUT/screen_results.json -o OUT/stage6_screened.json

Autonomous fallback (no agent)

python scripts/orchestrator.py <doc> --output-dir OUT

Runs every stage end-to-end as subprocesses; the reasoning stages call the Anthropic API on Sonnet (DeepSeek for dedup). The keyless free index search (Stage 4e) runs here by default (--no-freesearch to skip); web search (Stage 4d) is agent-only and not available in this runner. Use it for unattended runs or when no agent is driving. It is the fallback, not the default.

Undermind (Stage 1)

Undermind runs automatically from the extracted brief. The driver (undermind-search/scripts/undermind_search.py) launches Playwright, signs in with the credentials in ~/.lit-review-pipeline.env, drives the Classic search (sidebar Classic → Search → brief → the agent answers the clarifying questions → Generate Research Report), waits for the report, and exports the references (BibTeX by default). undermind_ingest.py then parses and enriches them into stage1_undermind.json (+ .bib). First-time setup stores the credentials with python undermind-search/scripts/undermind_search.py --login.

If credentials are missing or login/export fails, the stage degrades gracefully: it prints UNDERMIND_DEFERRED, writes empty results, and the pipeline continues on Google Scholar (the brief remains in undermind_brief.txt). When the Undermind UI changes, re-confirm the locators by driving the live site with the windows-mcp desktop tools, then update them in the driver (see undermind-search/SKILL.md). The previous Playwright automation for older UIs is preserved under _legacy/.

Scholar Labs (Stage 2)

Google Scholar Labs is opt-in (pass --scholarlabs; off by default because Google rate-limits its Cite/BibTeX export under automation, so it frequently defers — Undermind and Deep Research are the dependable deep-search channels). When enabled it runs from the Scholar Labs question Stage 0 writes to scholarlabs_query.txt — a single detailed research question, which is a different input from the Undermind brief. The driver (scholarlabs-search/scripts/scholarlabs_search.py) drives Playwright: it signs in to Google with a persistent profile (~/.scholar-profile), submits the question, waits for the result cards, and reads each result's citation through the standard Scholar Cite → BibTeX export; scholarlabs_ingest.py parses and enriches them into stage2_scholarlabs.json (+ .bib). It returns ~10 highly relevant papers that complement Undermind's broader set. First-time setup seeds the Google session with python scholarlabs-search/scripts/scholarlabs_search.py --login (sign in by hand, including any 2FA).

This stage runs off-screen, not headless. Google serves headless/automated Scholar an "unusual traffic" CAPTCHA, but a real headed Chrome passes — so the orchestrator runs it with --hidden: a genuine headed window positioned far off-screen, reusing the seeded session, so the stage works without anything appearing in front of you. If the session lapses and a 2FA wall appears (which an off-screen window can't clear), the stage defers with a desktop alert telling you to re-run --login (visible) to re-authorize. A missing sign-in or CAPTCHA also degrades gracefully (SCHOLARLABS_DEFERRED, empty results) and the pipeline continues. Disable it with --skip scholarlabs. See scholarlabs-search/SKILL.md.

Input parsing

scripts/manuscript_parser.py is bundled and ships with the skill, so a fresh clone works with no dependency beyond python-docx:

• .docx — the body is walked in document order so paragraphs and tables are captured, and tracked-change text is included (technique adopted from the word-docx skill's extract_text). Footnotes and endnotes are read directly from the document XML so they are never lost.
• .tex — title, abstract, sections, and \footnote{} content are extracted; the bibliography and \cite/\ref keys are stripped so they do not pollute the topic profile. \input/\include targets are inlined.

Short inputs such as a bare abstract are handled: the text becomes a single block and the extractor infers the framing.

API keys

Stored in ~/.lit-review-pipeline.env (auto-loaded; template in lit-review-pipeline.env.example). Real environment variables take precedence.

Variable	Needed for
ANTHROPIC_API_KEY	Autonomous fallback only (orchestrator.py): Stage 0, dedup, screening, and Undermind clarifying answers. The agent-driven flow needs no Anthropic key.
SEARCHAPI_API_KEY	Google Scholar + SSRN/HeinOnline/forthcoming (required for search)
GEMINI_API_KEY	Stage 2b Gemini Deep Research (default-on alternative deep search)
DEEPSEEK_API_KEY	LLM fuzzy dedup + Crossref title matching (recommended)
OPENALEX_API_KEY	OpenAlex Premium for enrichment (optional)
UNDERMIND_EMAIL / UNDERMIND_PASSWORD	Undermind login (Stage 1 driver; set on first run via --login)
SCHOLAR_EMAIL / SCHOLAR_PASSWORD	Google login for Scholar Labs (Stage 2 driver; set on first run via --login)

CLI flags

Flag	Default	Description
document	—	Path to a .tex/.docx document (positional)
--query	—	Run from a raw query instead of a document (skips Stage 0)
--output-dir	./lit-review-output	Output directory
--model	claude-sonnet-4-6	Model for Stage 0 extraction
--screen-model	claude-sonnet-4-6	Model for Stage 6 screening
--max-chars	30000	Max document characters sent to the extractor
--skip	none	Skip default stages: undermind, deepresearch, scholar, dedup, screen
--quick	off	Fast run: SearchAPI Google Scholar only (skips the Undermind and Deep Research deep-search stages)
--scholarlabs	off	Opt in to the Google Scholar Labs deep search (Stage 2); off by default because Google rate-limits its Cite export under automation
--ssrn --nber --heinonline --forthcoming	off	Opt-in supplementary sources
--citation-chain	off	Opt-in Semantic Scholar citation chaining
--no-llm	off	DOI-only dedup (skip LLM pass)
--verify / --no-verify	on	Cross-check papers vs OpenAlex/Crossref/Semantic Scholar and drop those none can confirm (anti-hallucination); --no-verify keeps all (dropped papers saved to stage5_merged_unverified.json)
--top-seeds	20	Seeds for citation chaining

Output

{output-dir}/
├── search_plan.json / .md          # research question, Undermind brief, queries, themes
├── scholar_queries.json            # the Google Scholar query list
├── undermind_brief.txt             # the Undermind brief
├── scholarlabs_query.txt           # the Scholar Labs research question
├── stage1_undermind.json / .bib    # Undermind results (enriched)
├── stage2_scholarlabs.json / .bib  # Scholar Labs results (enriched)
├── stage2b_deepresearch.json / .bib # Gemini Deep Research results
├── stage4a_scholar.json / .ris
├── stage5_merged.json / .ris       # deduplicated, verified master list
├── stage5_merged_unverified.json   # papers dropped by verification (audit trail)
├── stage6_screened.json/.xlsx/.ris        # ALL screened papers, scored & ranked (no score filter)
├── stage6_screened.bib                     # screened papers (BibTeX), score >= 4 only
├── stage6_filtered.json/.xlsx              # score >= 4 subset (the shortlist)
├── dedup_log.json / dedup_report.md / verification.log
└── pipeline.log

Skill structure

lit-review-orchestrator/
├── SKILL.md
├── AGENTS.md                         # Codex repository guidance
├── CLAUDE.md                         # Claude Code repository guidance
├── agents/openai.yaml                # Codex UI metadata and explicit invocation policy
├── docs/
│   ├── codex.md                      # Codex model routing and subagent rules
│   ├── claude-code.md                # Claude Code model routing and tool notes
│   └── images/                       # diagram sources and rendered images
├── requirements.txt
├── lit-review-pipeline.env.example
├── scripts/
│   ├── orchestrator.py            # this controller
│   ├── lit_review_gui.py          # interactive settings dialog (GUI front door)
│   ├── manuscript_parser.py       # bundled .docx/.tex parser
│   └── extract_search_plan.py     # Stage 0 extraction
├── undermind-search/              # Stage 1 (Playwright driver + ingest)
├── scholarlabs-search/            # Stage 2 (Playwright driver + ingest)
├── deepresearch-search/           # Stage 2b (Gemini Deep Research API)
├── supplementary-search/          # Google Scholar + opt-in sources
├── lit-dedup/                     # Stage 5
├── lit-screen/                    # Stage 6
├── examples/sample_manuscript.tex
└── _legacy/                       # previous Undermind Playwright automation

Operational notes

• Run in the foreground; the dashboard streams stage status to the console and pipeline.log.
• Stages 5 (dedup) and 6 (screen) can be re-run standalone after editing any stage input file.
• Citation chaining needs DOI-bearing seeds, so it is most useful once Undermind enrichment is active.
• Stage 0 failure aborts the run (the pipeline cannot proceed without a search plan).
• The SSRN, HeinOnline, and forthcoming sources are Google Scholar searches with a site:/source: filter — subsets of Stage 4a — so they mainly force those venues to surface rather than adding a new index. NBER (nber.org API) and citation chaining (Semantic Scholar) hit independent indexes and add genuine coverage.
• --quick runs only the SearchAPI Google Scholar channel (no browser, no login), useful for a fast pass or when the Undermind/Scholar Labs logins are unavailable.