collaboration_skill

This skill guides teams to manage Claude Code collaboration centrally, standardize knowledge, and optimize workflows across projects.
  • Shell

0

GitHub Stars

1

Bundled Files

2 months ago

Catalog Refreshed

4 months ago

First Indexed

Readme & install

Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.

Installation

Preview and clipboard use veilstrat where the catalogue uses aiagentskills.

npx veilstrat add skill delphine-l/claude_global --skill collaboration

  • SKILL.md27.6 KB

Overview

This skill describes best practices for using Claude Code in team environments, focusing on centralized skill management, knowledge capture, version control, and collaborative workflows. It explains how to organize shared skills in a single metadata location and how teams should symlink and maintain those skills for consistency. The goal is predictable, auditable, and discoverable team knowledge.

How this skill works

At session start Claude reads skill files linked into the project directory and uses them as context for guidance. Skills persist as files in a central metadata location ($CLAUDE_METADATA) and are not modified by session conversation unless a user explicitly updates them. Teams maintain and version-control the central metadata so updates propagate to all projects that use symlinks to those skills.

When to use it

  • Onboard a new project: symlink only the skills needed for that project from $CLAUDE_METADATA.
  • After discovering a recurring fix or optimization that benefits the team.
  • When standardizing workflows, tool flags, or configuration patterns.
  • Before major releases or audits to ensure documented team practices reflect current reality.
  • During periodic reviews to consolidate fragmented knowledge and remove one-off tips.

Best practices

  • Keep all skill files in the central metadata location ($CLAUDE_METADATA); never store canonical skill content inside individual project folders.
  • Use symlinks from projects to the centralized skills so updates apply everywhere without duplication.
  • Version-control the central metadata and require review for significant changes to preserve auditability.
  • Capture only team-relevant, repeatable knowledge; avoid adding one-off or personal-environment fixes.
  • Split long guidance into a short summary plus supporting documents to limit what Claude loads by default.

Example use cases

  • Document a repeated HPC workflow timeout fix as a troubleshooting entry so all team members can apply it.
  • Add a token-efficiency guideline (e.g., default quiet mode flags) after measuring cost savings.
  • Set up a new project by symlinking domain-specific skills and shared commands from $CLAUDE_METADATA.
  • Run an end-of-session summary and ask Claude to suggest candidate updates, then review and commit the changes.
  • Perform a monthly audit of skills to merge duplicates, remove stale tips, and improve discoverability.

FAQ

Centralizing skills ensures a single source of truth, simplifies updates, and enables version control and team discovery.

Can Claude auto-update skills from a session?

No. Session learnings are temporary; skill files persist only when a user explicitly edits and commits updates in the central metadata.

Built by
VeilStrat
AI signals for GTM teams
© 2026 VeilStrat. All rights reserved.All systems operational