Skills Reference

Skills Reference

Every skill documented with its arguments, integrations, and examples.

---

/ops — Command center router

File: skills/ops/SKILL.md
Description: Routes to the right sub-skill based on natural language input. If called with no arguments, shows an interactive menu.

Arguments: [command] [args]

Input	Routes to
(empty), go, morning, briefing	/ops:go
setup, configure, init, install	/ops:setup
inbox, unread, messages	/ops:inbox
comms, send, whatsapp, email, slack, telegram	/ops:comms
fires, incidents, down, sentry	/ops:fires
projects, dashboard, status	/ops:projects
next, priority, what	/ops:next
triage, issues	/ops:triage
linear, sprint, board	/ops:linear
revenue, money, mrr, costs	/ops:revenue
deploy, ship	/ops:deploy
merge, prs	/ops:merge
yolo	/ops:yolo

Example:

/ops morning
/ops send "shipping today" to team on Slack

---

/ops:setup — Interactive setup wizard

File: skills/setup/SKILL.md
Description: Walks through every configuration step interactively. Installs CLIs, connects channels, builds the project registry, saves preferences. Run once after install or any time to reconfigure.

Arguments: [section]

Sections:

• cli — Install missing command-line tools
• channels — Configure Telegram, WhatsApp, Email, Slack
• mcp — Connect Linear, Sentry, Vercel, Gmail via OAuth
• plugins — Install companion plugins (GSD)
• registry — Build the project registry
• prefs — Set owner name, timezone, default channels
• env — Export CLAUDE_PLUGIN_ROOT in shell profile

Examples:

/ops:setup
/ops:setup telegram
/ops:setup registry

Integrations used: Bash, Read, Write, AskUserQuestion
Data written to:

• ~/.claude/plugins/data/ops-ops-marketplace/preferences.json — credentials and preferences
• scripts/registry.json — project registry (gitignored)

---

/ops:go — Morning briefing

File: skills/ops-go/SKILL.md
Description: Pre-gathers all business data via shell scripts in parallel, then presents a unified dashboard. The primary daily driver.

Arguments: [project-alias]

Data gathered:

• Infrastructure: ECS cluster health via ops-infra
• Git status: branch state, dirty files, ahead/behind via ops-git
• Open PRs: count, review state, CI status via ops-prs
• CI failures: last 24h across all repos via ops-ci
• Unread messages: WhatsApp, Email, Slack, Telegram counts via ops-unread
• GSD state: active phases and progress (if GSD installed)
• Calendar: today's events via gog or Google Calendar MCP

Output format:

FIRES     [critical issues or "none"]
COMMS     [unread counts per channel]
PRs       [open PRs with status]
CI        [pass/fail per repo]
SPRINT    [Linear sprint summary]
REVENUE   [AWS spend this month]
CALENDAR  [today's events]

Integrations used: gh, aws, gog, Slack MCP, Telegram MCP

Example:

/ops:go
/ops:go myapp    # scope to one project

---

/ops:inbox — Unified inbox

File: skills/ops-inbox/SKILL.md
Description: Full inbox management across all channels. Scans the complete inbox (not just unread), classifies conversations, drafts replies, archives handled threads.

Arguments: [channel: whatsapp|email|slack|telegram|all]

Per-channel behavior:

Channel	Read	Send	Archive	Requires
WhatsApp	All chats (7 day window)	Yes	No (wacli limitation)	wacli
Email	Full inbox (max 30 threads)	Yes	Yes (via gog gmail archive)	gog CLI
Slack	MCP search, read channels	Yes	No	Slack MCP
Telegram	All dialogs	Yes	No	Telegram MCP

Classification: Each conversation is tagged as NEEDS_REPLY, WAITING (on other party), or HANDLED.

Examples:

/ops:inbox
/ops:inbox email
/ops:inbox whatsapp
/ops:inbox slack

Integrations used: wacli, gog, Gmail MCP, Slack MCP, Telegram MCP

---

/ops:comms — Cross-channel messaging

File: skills/ops-comms/SKILL.md
Description: Send and read messages across all channels. Routes based on natural language or explicit channel specification.

Arguments: [channel] | send [message] to [contact] | read [channel]

Examples:

/ops:comms send "shipping today" to John Smith
/ops:comms send "build is green" to #deploys on Slack
/ops:comms read whatsapp
/ops:comms read telegram

Integrations used: wacli, Gmail MCP, Slack MCP (slack_send_message, slack_read_channel, slack_search_users), Telegram MCP (send_message, get_updates)

---

/ops:fires — Production incidents dashboard

File: skills/ops-fires/SKILL.md
Description: Reads ECS health, Sentry errors, and CI failures. Classifies by severity. Offers to dispatch fix agents.

Arguments: [project-alias|all]

Severity levels:

Level	Criteria
CRITICAL	Service down, DB unreachable, auth broken
HIGH	Elevated error rate, deploy stuck, CI main broken
MEDIUM	Non-critical service degraded, flaky tests
LOW	Warning-level, non-urgent

Actions offered:

• Dispatch triage agent for specific issue
• View ECS logs for a service
• Open Sentry dashboard link

Integrations used: aws (ECS), gh (Actions), Sentry MCP

Examples:

/ops:fires
/ops:fires myapp

---

/ops:deploy — Deploy status

File: skills/ops-deploy/SKILL.md
Description: Shows ECS service versions, Vercel deployments, recent deploys, pending deploys, and CI/CD pipeline state across all projects.

Arguments: [project-alias|ecs|vercel|all]

Integrations used: aws (ECS describe-services), Vercel MCP (list_deployments, list_projects, get_deployment, get_runtime_logs, get_deployment_build_logs)

Examples:

/ops:deploy
/ops:deploy vercel
/ops:deploy myapp

---

/ops:revenue — Revenue and costs

File: skills/ops-revenue/SKILL.md
Description: AWS spend tracking, credits balance, project revenue stages, burn rate, and runway estimate.

Arguments: [costs|revenue|credits|runway|all]

Data gathered:

• Current month costs by service (AWS Cost Explorer)
• Last 3 months trend
• AWS credits balance and expiry dates
• Project revenue stages from registry.json

Integrations used: aws (Cost Explorer: get-cost-and-usage, list-cost-allocation-tags)

Examples:

/ops:revenue
/ops:revenue costs
/ops:revenue credits
/ops:revenue runway

---

/ops:projects — Portfolio dashboard

File: skills/ops-projects/SKILL.md
Description: Shows all registered projects with GSD phase, current branch, uncommitted changes, CI status, and suggested next action. Jump to any project by alias.

Arguments: [project-alias]

Per-project data shown:

• Active branch and dirty file count
• Commits ahead/behind remote
• Open PR count with review state
• CI status (last run result)
• GSD phase and progress (if GSD installed)
• Next suggested action

Integrations used: git, gh, GSD .planning/STATE.md

Examples:

/ops:projects
/ops:projects myapp

---

/ops:linear — Linear sprint management

File: skills/ops-linear/SKILL.md
Description: Linear command center. Shows current sprint, creates and updates issues, manages priorities, syncs with GSD phases.

Arguments: [sprint|create|update|sync|backlog|issue-id]

Linear MCP tools used: list_issues, list_cycles, list_teams, list_projects, get_issue, save_issue, get_team, list_users, save_comment

Examples:

/ops:linear
/ops:linear sprint
/ops:linear create "Fix auth token refresh"
/ops:linear HEA-42
/ops:linear backlog

---

/ops:triage — Cross-platform issue triage

File: skills/ops-triage/SKILL.md
Description: Pulls issues from Sentry, Linear, and GitHub. Cross-references against code to find already-fixed issues. Auto-resolves stale ones. Dispatches triage agents for active issues.

Arguments: [project-alias|sentry|linear|github|all]

Triage process:

1. Pull issues from all configured sources
2. Search codebase for error patterns to check if fixed
3. Auto-resolve issues where the fix is already in main
4. For active issues, offer to dispatch triage-agent per issue

Integrations used: Sentry MCP (search_issues, update_issue), Linear MCP (list_issues, save_issue), gh (issue list)

Examples:

/ops:triage
/ops:triage sentry
/ops:triage myapp

---

/ops:merge — Autonomous PR merge pipeline

File: skills/ops-merge/SKILL.md
Description: Scans all repos for open PRs. Dispatches subagents to fix CI, resolve conflicts, address review comments, then merges. The most powerful automation in the plugin.

Arguments: [--main] [--repo org/repo] [--dry-run]

Flags:

• --main — Also sync dev→main branches after merging feature PRs
• --repo org/repo — Scope to a single repository
• --dry-run — Preview what would happen without making changes

Pipeline per PR:

1. Check CI status and review state
2. If CI failing: read logs, fix code, push fix commit
3. If merge conflict: resolve conflict, push
4. If review comments: address comments in code
5. Once green and approved: merge

Pre-gathering: Uses ops-merge-scan bin script to collect all PR data with CI state, mergeable flag, and review decisions before spawning agents.

Integrations used: gh (PR list, merge, CI logs, reviews), Git, Read/Edit/Write for code fixes

Examples:

/ops:merge
/ops:merge --dry-run
/ops:merge --repo org/myapp
/ops:merge --main

---

/ops:next — Next action advisor

File: skills/ops-next/SKILL.md
Description: Applies the business priority stack and tells you the single most important thing to do next. Routes to the right skill.

Arguments: [context]

Priority stack:

1. Production fires (CRITICAL/HIGH from /ops:fires)
2. Urgent communications (unread messages from key contacts)
3. Ready-to-merge PRs (green CI, approved)
4. Linear sprint work (in-progress issues)
5. Revenue-generating GSD work (from roadmap)

Examples:

/ops:next
/ops:next focus on myapp

Integrations used: ops-infra, ops-prs, ops-unread, Linear MCP, Slack MCP

---

/ops:yolo — YOLO mode

File: skills/ops-yolo/SKILL.md
Description: Spawns 4 parallel C-suite agents. Each analyzes the business from their perspective. Produces an unfiltered Hard Truths report. After typing YOLO, autonomously runs the business.

Arguments: [YOLO|analyze|report]

Phase 1 — Pre-gather all data: All shell scripts run simultaneously (infra, git, PRs, CI, unread, AWS costs, registry, GSD state).

Phase 2 — Spawn 4 agents in parallel:

Agent	Model	Focus
CEO (yolo-ceo.md)	claude-opus-4-6	Strategic synthesis — reads and resolves CTO/CFO/COO reports
CTO (yolo-cto.md)	claude-opus-4-6	Tech debt, production risks, scalability limits
CFO (yolo-cfo.md)	claude-opus-4-6	Burn rate, waste, runway, ROI
COO (yolo-coo.md)	claude-opus-4-6	Execution gaps, stale work, broken processes

Phase 3 — After typing YOLO: Enters autonomous loop: inbox → merge → fires → GSD phases → deploys.

Integrations used: All available integrations. Linear MCP, Vercel MCP, Slack MCP, Gmail MCP, aws, gh, GSD.

Examples:

/ops:yolo
/ops:yolo analyze
# After reading report:
YOLO