How do I install the Write Paper skill?

Run npx skillmds@latest add emmanuelrtm/write-paper in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Write Paper safe to use?

This skill has not completed SkillMD's automated safety review yet. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Write Paper?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Write Paper free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Write Paper?

EmmanuelRTM (@emmanuelrtm) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Write Paper by emmanuelrtm — Agent Skill

Name: Write Paper
Author: EmmanuelRTM

Write Paper

Run (or re-run) Stage 3 only — the Paper Writer pipeline: CONCEIVE (LLM) → GROUND (deterministic script) → CRITIC (fresh-context LLM) → RESOLVE+COMPOSE (LLM) producing the evidence-tagged paper draft. Requires best/SELECTED.json from /discover.

EmmanuelRTM 93d0538 1.9 KB Updated Sep 17, 2026

File contents

/write-paper — Stage 3: CONCEIVE → GROUND → CRITIC → RESOLVE → COMPOSE

Arguments: $ARGUMENTS

Current state: !python3 .claude/scripts/ledger.py status

Preconditions: resolve the run (--run or ACTIVE_RUN); best/SELECTED.json and best/eval.json must exist — otherwise stop and point to /discover.

CONCEIVE: launch paper-writer (stage=conceive, run dir). Expect paper/research-representation.md, fully evidence-tagged.
GROUND: python3 .claude/scripts/ground_check.py <run>.
- Exit 1 (grounding ratio < 0.85): relaunch paper-writer (stage=conceive) once with instruction "fix every non-SUPPORTED check in paper/ground-report.json". Re-run ground_check; if still failing, stop and show the unsupported list to the user.
CRITIC: launch paper-critic (run dir). Expect paper/critic-report.md.
RESOLVE + COMPOSE: launch paper-writer (stage=resolve-compose, run dir). Expect paper/draft.md with all tags intact.
Convergence round (the paper's Ground→Critic→Resolve loop, max 2 rounds, terminating on convergence — zero blockers — or plateau): if step 3's report contained any BLOCKER, relaunch paper-critic against paper/draft.md. If BLOCKERs remain, relaunch paper-writer (stage=resolve-compose) once more with them; then stop regardless and report residual issues honestly.
Ledger: python3 .claude/scripts/ledger.py append '{"event":"stage_write_paper","detail":"draft complete, ratio=<r>"}'

Report: grounding ratio, critic issue count (blockers), draft section list. Next stage: /verify-claims.

Frequently asked questions