# Literature Scout

> Literature Scout

- Skill: `sapientinc/literature-scout-2` (Agent Skill, multi-file: 3 files)
- Install (CLI): `npx skillmds@latest add sapientinc/literature-scout-2`
- Raw SKILL.md: https://api.skillmd.com/api/skills/sapientinc/literature-scout-2/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Research & Search
- Author: sapientinc (https://skillmd.com/u/sapientinc)
- Updated: 2026-09-22
- Page: https://skillmd.com/skills/sapientinc/literature-scout-2

---

# Literature Scout

Check whether a toy conjecture or proof sketch duplicates task-local prior work.

Hard boundaries:
- This role is disabled by default in smoke tests.
- Keep conjecture-specific search policy and reading notes inside this task
  project.
- Use public lookup only through the standard task-approved literature tool set;
  otherwise rely on task-local prior work.
- If a source mentions external data, software, dependency installs, or a new
  runtime environment, do not acquire it. Adapt the idea to the toy task's
  existing local harness and record the missing resource only as context.
- Return normalized literature, public database, open-access provenance, or
  prior-art evidence only; do not decide the research loop outcome.