Claude Multi Session Research Skills — Development skill for Claude Code
Personal Claude Code skills for running work across several concurrent sessions — one that builds, one that explains, one that audits — and for keeping experimental results trustworthy enough to cite.
How to install Claude Multi Session Research Skills
This entry records only its repository, not the path inside it, so there is no
exact command to give. Open HelloJJJie/claude-multi-session-research-skills and copy the folder into
~/.claude/skills/, or the file into ~/.claude/agents/.
What Claude Multi Session Research Skills does
Personal Claude Code skills for running work across several concurrent sessions — one that builds, one that explains, one that audits — and for keeping experimental results trustworthy enough to cite later.
Alternatives in Development
- Bro Skill — 🤙 /bro — re-explains the last answer in plain language, for when the reply made you go 'bro what' 296 ★
- Statistics Page Writer — Deep-research and publish a sourced statistics page journalists cite for months 195 ★
- Scipilot Cite Skill — SciPilot family - Academic citation discovery & insertion Skill for Claude Code /codex 84 ★
README
Claude Code skills
A personal skill set for running work across several Claude Code sessions — one that builds, one that explains, one that audits — and for keeping experimental results trustworthy enough to cite months later.
These are **not** a framework. Every rule here exists because a specific mistake got through, and each one carries the failure that produced it. They are written to be edited: see `skill-maintenance`.
Two layers
Skills here are **generic** — they should help a project with entirely different subject matter. Anything that needs a particular project's nouns belongs in that project's own skill (`/.claude/skills//`), not here.
| skill | what it owns |
|---|---|
multi-session-protocol |
Roles; who owns a claim and its scope; who may close a finding; how a retraction propagates instead of staying alive in documents nobody re-read. Load first when sessions share a project. |
experiment-engineering |
The build role. Your own pipeline is the first suspect behind a result; decomposing a null against a published configuration; long jobs that survive interruption. |
experiment-provenance |
Append-only run directories; what a run's metadata must answer; when a result may not be cited; grading design decisions A/B/C; pre-registering before the run. |
research-companion |
The learn role. Explaining a codebase at the user's pace, graded literature work, re-deriving the builder's numbers before believing them. |
project-audit |
The audit role. Checking claims against the filesystem rather than against conversation. Never briefed by the sessions it audits. |
study-notes |
Turning a study session into atomic Obsidian notes with a linking index. |
publishing-to-github |
Publishing a repo and keeping it updated: the checks before an irreversible first push, and explaining every subsequent one — what changed, why, and what a user should do differently. |
skill-maintenance |
Keepin |
Related Skills
Skill Anthropic Grade Optimizer
Audits Claude-directing artifacts (CLAUDE.md, SKILL.md, subagent, hook, MCP, prompts, api_config) against 189
Claude Public
A public template of my personal Claude Code configuration: settings, sync scripts, file-manager agent, and a
Claude Short Answer Skill
Initiate with /short - A Claude Code skill that switches output to summary-only mode - cutting reasoning, walk
SwarmBar
SwarmBar: a macOS menu bar app showing live status of CLI coding agents (Claude Code, Codex, Kimi Code, BearCo
Brain Reconcile
Find and resolve contradictions across the second brain, keeping a history trail. Never edits preserved or raw
Claude Account Sync Skill
A Claude Code skill that keeps the Claude desktop app's Code sidebar identical across several Claude accounts
Related Agents
Data Lead
L0.9 Data team lead — owns the measurement strategy for a product: which metrics matter, how they're defined,
2code
AI-assisted development produces a mess of parallel state: several terminals, experimental branches, uncommitt
Research Paper Writer
Use this agent when you need to write academic papers, research reports, or technical manuscripts based on exp