| name | researchstack-literature-map |
| description | Related-work and positioning skill for CS, networking, and AI papers. Use when Codex needs to map prior work,
identify the closest baselines and competing narratives, clarify novelty boundaries, or build a literature
matrix that can support idea review, experiment design, or paper writing.
|
Literature Map
Read ../references/novelty-boundary.md.
Build a compact literature matrix, not a generic survey.
For each relevant prior work, capture:
- problem setting,
- method or mechanism,
- assumptions,
- evaluation setting,
- strongest result,
- limitation most relevant to the current paper.
Group papers by competing explanation, not by publication year alone.
Always identify:
- The two or three papers reviewers will treat as the closest comparison.
- The line of work that makes the new idea look incremental.
- The line of work that the paper can honestly claim to extend or unify.
- The strongest industrial/open-source baseline reviewers will expect.
- The foundational predecessor that introduced the core architecture, metric, or mechanism.
When writing positioning guidance, produce:
- a novelty boundary statement,
- a novelty-boundary label:
clear gap, crowded but viable, incremental, already done, or unknown,
- a closest-work matrix,
- a baseline shortlist,
- likely reviewer objections rooted in prior work,
- story candidate: likely venue story pattern, X/Y/Z statement if systems, closest-work contrast sentence, and first decisive figure or experiment,
- related-work paragraph structure.
If evidence is insufficient to claim novelty, say so directly.
If current literature is fast-moving or the user asks whether an idea has been done, browse/search. Do not rely only on memory for novelty claims.