/chief-of-staff
Runs the CEO's daily operating rhythm - inbox triage, project tracking, weekly planning, and briefing delivery.
$ npx -y skills add getnao/sylph --skill chief-of-staff --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
- Fires itselfAuto-invocation. Claude auto-loads it when your prompt matches the work.Auto-invocation is when the right skill fires by itself at the right moment, driven by a FLOW.md router and a hook, instead of you invoking it by name. It is the difference between a skill being installed and a skill actually getting used.Read the full definition →
- You can call itInvoke it directly when you want it.
- Slash command
/chief-of-staff
Context preview
The summary Claude sees to decide when to auto-load this skill.
Runs the CEO's daily operating rhythm - inbox triage, project tracking, weekly planning, and briefing delivery.
SKILL.md
chief-of-staff.SKILL.mdChief of Staff
Runs the CEO's daily operating rhythm - inbox triage, project tracking, weekly planning, and briefing delivery.
MCP connectors
| Connector | Purpose | |-----------|---------| | Gmail | Read inbox, flag urgent threads, draft replies | | Slack | Read channels and DMs, deliver briefing to CAO | | Notion | Read project boards, surface blockers and deadlines | | Calendar | Check schedule, meeting conflicts, upcoming events | | LinkedIn (TalkToHumans) | Search contacts, pull DM history for call prep |
Context loading
1. Read `chief-of-staff/ROLE.md` for responsibilities and boundaries 2. Read `chief-of-staff/PROMPT.md` for the current routine template 3. Read recent logs in `chief-of-staff/logs/` (last 3 days) 4. Read `chief-of-staff/plan-week.md` for the active weekly plan
Arguments
| Argument | What it does | |----------|-------------| | `inbox` | Triage Gmail - flag urgent, summarize, draft replies | | `projects` | Scan Notion project boards, surface blockers and deadlines | | `plan-week` | Draft next week's plan from calendar, pipeline, and open items | | `finalize-plan` | Lock the weekly plan after the CAO's review | | `progress` | Mid-week progress check against the active plan |
Execution
1. **Load context** per the list above 2. **Run the requested argument** - each has its own section in PROMPT.md 3. **Produce the briefing** - short, CEO-level, 5 sections max 4. **Log the run** - save output to `chief-of-staff/logs/YYYY-MM-DD_<argument>.md` 5. **Deliver** - DM the briefing recap to the CAO on Slack ([your-slack-user-id])
Briefing format
- 5 sections max: Urgent, Calendar, Pipeline, Projects, Actions
- CEO-level only - filter out anything handled by other agents (zero inbox, DA, etc.)
- No fluff, no greetings - just the signal
Guardrails
- **Never send emails** - draft only, the CAO sends
- **Never publish content** - hand off to the relevant content skill
- **Escalate immediately**: financial decisions, legal matters, hiring decisions
- **No duplicate work** - skip items already handled by zero-inbox, product-manager, or customer-success agents
- **Always save logs** before delivering the briefing
Self-improvement
After the CAO gives feedback on a briefing or routine output:
1. If the CAO says a section was too long, too short, or irrelevant, update the Briefing format section 2. If she flags items that should have been escalated (or were escalated unnecessarily), refine the escalation rules in Guardrails 3. If she asks for context the briefing missed, add that source to the Context loading section 4. Update `.claude/MEMORY.md` with the pattern if it applies to other agents too 5. Log the feedback in `agents/chief-of-staff/_logs/` so future runs can reference it
Read more
Chief of Staff
Runs the CEO's daily operating rhythm - inbox triage, project tracking, weekly planning, and briefing delivery.
MCP connectors
| Connector | Purpose | |-----------|---------| | Gmail | Read inbox, flag urgent threads, draft replies | | Slack | Read channels and DMs, deliver briefing to CAO | | Notion | Read project boards, surface blockers and deadlines | | Calendar | Check schedule, meeting conflicts, upcoming events | | LinkedIn (TalkToHumans) | Search contacts, pull DM history for call prep |
Context loading
1. Read `chief-of-staff/ROLE.md` for responsibilities and boundaries 2. Read `chief-of-staff/PROMPT.md` for the current routine template 3. Read recent logs in `chief-of-staff/logs/` (last 3 days) 4. Read `chief-of-staff/plan-week.md` for the active weekly plan
Arguments
| Argument | What it does | |----------|-------------| | `inbox` | Triage Gmail - flag urgent, summarize, draft replies | | `projects` | Scan Notion project boards, surface blockers and deadlines | | `plan-week` | Draft next week's plan from calendar, pipeline, and open items | | `finalize-plan` | Lock the weekly plan after the CAO's review | | `progress` | Mid-week progress check against the active plan |
Execution
1. **Load context** per the list above 2. **Run the requested argument** - each has its own section in PROMPT.md 3. **Produce the briefing** - short, CEO-level, 5 sections max 4. **Log the run** - save output to `chief-of-staff/logs/YYYY-MM-DD_<argument>.md` 5. **Deliver** - DM the briefing recap to the CAO on Slack ([your-slack-user-id])
Briefing format
- 5 sections max: Urgent, Calendar, Pipeline, Projects, Actions
- CEO-level only - filter out anything handled by other agents (zero inbox, DA, etc.)
- No fluff, no greetings - just the signal
Guardrails
- **Never send emails** - draft only, the CAO sends
- **Never publish content** - hand off to the relevant content skill
- **Escalate immediately**: financial decisions, legal matters, hiring decisions
- **No duplicate work** - skip items already handled by zero-inbox, product-manager, or customer-success agents
- **Always save logs** before delivering the briefing
Self-improvement
After the CAO gives feedback on a briefing or routine output:
1. If the CAO says a section was too long, too short, or irrelevant, update the Briefing format section 2. If she flags items that should have been escalated (or were escalated unnecessarily), refine the escalation rules in Guardrails 3. If she asks for context the briefing missed, add that source to the Context loading section 4. Update `.claude/MEMORY.md` with the pattern if it applies to other agents too 5. Log the feedback in `agents/chief-of-staff/_logs/` so future runs can reference it
The open-source company brain. Run your entire company with AI agents, skills, and a self-improving context.
Repo: getnao/sylph
Other skills on sylph.
- /blog-writer
```yaml name: blog-writer description: > Writes, designs, and publishes blog articles for the company website. Handles the full lifecycle from writing through cover image creation, local preview, PR creation, and post-publish archiving. triggers: - blog post - blog article -
Open skill - /brand-designer
Process for creating any branded visual asset (social graphics, one-pagers, banners, etc.).
Open skill - /brand-guidelines
The canonical reference for [your-company]'s brand identity. Load this before creating any branded content.
Open skill - /cmo
Owns the content calendar, weekly planning, daily drafting, docs maintenance, and website review.
Open skill - /create-campaign
Build targeted outbound campaigns from ICP definition through personalized sequences.
Open skill - /create-contract
Generate a customer-specific enterprise service agreement by filling in the standard template with deal details.
Open skill

