/paper-trail-new-sota

description: Create a new State-of-the-Art review on a topic, guaranteed without hallucinated citations. Inverted workflow (research → acquire PDFs → read → write), refuses to write from memory.

`/paper-trail:new-sota` — Create a SOTA on a topic

Invoke the `sota-writer` skill to produce a new SOTA / literature review on a given topic.

Usage

/paper-trail:new-sota <topic>
/paper-trail:new-sota <topic> --depth shallow|standard|exhaustive
/paper-trail:new-sota <topic> --max-candidates 30

What it does

0. **Pré-vol** : invoke `python3 -m pipeline preflight` to verify the environment (vault path, Python deps, `paper-search` MCP registered). If pre-flight reports `paper-search MCP non enregistré`, **halt and show the install recipe printed by preflight** before going further. The user must install the MCP, restart Claude Code, then re-run. 1. **Phase A — Research** : invokes `researcher` sub-agent for multi-source academic search (paper-search MCP across 22 platforms ; optionally NotebookLM + WebSearch) 2. **Phase B — Acquisition** : pushes candidates into the registry with state `candidate`, then invokes `pdf-cascade` to download + page 1 validation 3. **Phase C — Reading** : for each ref in `page1_validated`, opens the PDF and writes structured notes in the ref's markdown body 4. **Phase D — Writing** : produces the SOTA at `adapter.sota_output_path(<topic_slug>)`, citing only refs in `page1_validated`+, with a separate « Refs écartées » section listing rejected candidates and reasons

Safeguards

• **Refuses to write from memory** : every claim must trace back to

a `page1_validated` ref's notes body

• **>30% DROP refusal** : if more than 30% of candidates fail to

reach `page1_validated`, refuses to write the SOTA and reports the situation (user must refine topic, enable shadow libs, or intervene manually on specific refs)

• **PreToolUse hook** : the plugin blocks the final SOTA write if

any cited wikilink points to a non-`page1_validated` ref (see `tools/precheck_sota_wikilinks.py`)

Shadow libraries

To maximize acquisition success, activate shadow libs **before** running new-sota (see `DISCLAIMER.md`) :

export RESEARCH_ENABLE_SHADOW_LIBS=1
/paper-trail:new-sota "Petri nets in music notation"

Without shadow libs, paywall papers go to `blocked_human:cascade_exhausted` and won't be cited in the SOTA (will appear in « Refs écartées »).

Output

• New SOTA markdown file at the path resolved by the adapter
• Each cited ref's frontmatter updated with `cited_in[]`
• Phase report (number of candidates, validated, dropped, rejected)
• Recap of next actions if any refs need human intervention

Examples

Standard usage

/paper-trail:new-sota "Petri nets in music notation"

Limited depth

/paper-trail:new-sota "GPT-style transformer for symbolic music" --max-candidates 15

With explicit domain hint

/paper-trail:new-sota "polyrhythmic structure quantification" --depth exhaustive

Cross-reference

• `/paper-trail:audit-sota <path>` to audit an existing SOTA
• `/paper-trail:cascade <slug>` to acquire a specific ref manually
• `pipeline/USAGE.md` for the underlying worker B CLI