$ man context-wiki/co-work-sessions
Modes and Workflowsintermediate
Claude Code Co-Work
Shared folder sessions that turn passive documentation into active playbooks
What Co-Work Sessions Are
A co-work session is a Claude Code instance pointed at a shared repo. Every team member gets the same context: CLAUDE.md rules, skills, data files, voice system. The AI reads your playbooks on session start. It does not just store them - it executes them.
Traditional team docs rot. Confluence pages written six months ago. Notion wikis nobody updates. Google Docs with conflicting versions. A co-work folder is different because the documentation is the automation. The skill file that describes how to research a prospect also runs that research. The workflow that explains enrichment also performs the enrichment.
Traditional team docs rot. Confluence pages written six months ago. Notion wikis nobody updates. Google Docs with conflicting versions. A co-work folder is different because the documentation is the automation. The skill file that describes how to research a prospect also runs that research. The workflow that explains enrichment also performs the enrichment.
PATTERN
How SDRs Should Be Working Now
Drop a folder with: CLAUDE.md (team rules, voice, guardrails), scripts/ (enrichment, scoring, campaign automation), skills/ (research, outreach, pipeline review), data/ (target accounts, enrichment results).
New SDR joins. Opens Claude Code in the folder. Says "research Acme Corp for outbound." The agent runs the research skill - pulls Exa intelligence, enriches through Apollo, checks Attio history, produces a brief. The SDR shipped work before reading a single onboarding doc.
This is the active playbook model. The repo is the onboarding. Context compounds across sessions through handoff files. Every session builds on the last.
New SDR joins. Opens Claude Code in the folder. Says "research Acme Corp for outbound." The agent runs the research skill - pulls Exa intelligence, enriches through Apollo, checks Attio history, produces a brief. The SDR shipped work before reading a single onboarding doc.
This is the active playbook model. The repo is the onboarding. Context compounds across sessions through handoff files. Every session builds on the last.
PRO TIP
Active Playbooks vs Passive Docs
Passive doc: "To research a prospect, check LinkedIn, look at funding, review tech stack on BuiltWith, summarize in a Google Doc."
Active playbook: a skill file Claude Code reads and executes. Checks LinkedIn via browser, pulls funding from Exa, queries BuiltWith through the API, writes the summary to the research folder. Same workflow. One requires humans to follow steps. The other requires humans to say "research this company."
Every passive doc in your org is a candidate for conversion. Identify the steps. Write them as a skill. Test. Deploy to the shared folder. The documentation becomes the automation. The gap between knowing what to do and doing it closes to zero.
Active playbook: a skill file Claude Code reads and executes. Checks LinkedIn via browser, pulls funding from Exa, queries BuiltWith through the API, writes the summary to the research folder. Same workflow. One requires humans to follow steps. The other requires humans to say "research this company."
Every passive doc in your org is a candidate for conversion. Identify the steps. Write them as a skill. Test. Deploy to the shared folder. The documentation becomes the automation. The gap between knowing what to do and doing it closes to zero.
knowledge guide
related entries