Rebuilt from scratch following Codex CLI's actual architecture (as outlined in the review on PR #26). Closes the integration gap with a proper build adapter, correct install paths, TOML agent files, and all the architectural differences documented and tested. What changed vs the previous attempt (PR #26): - Agents: build adapter generates .toml files (name/description/ developer_instructions) into dist/codex-cli/.codex/agents/ instead of copying .md files with sed transforms - Skills: installed to .agents/skills/ (correct Codex discovery path) instead of .codex/skills/ - Dispatcher: AGENTS.md uses a root-context orchestration header that works within agents.max_depth=1 constraints; named-agent routing replaced with embedded-instructions workaround for the known spawn_agents limitation (openai/codex#15250) - Tool compat: AskUserQuestion and request_user_input removed; all prompts adapted to Codex's actual tool set and approval/confirmation flow - Installer/updater: launchme.sh --platform codex-cli and updateme.sh with Codex auto-detection, creating the correct split layout (AGENTS.md + .codex/agents/ + .codex/config.toml + .agents/skills/) - Tests: new per-adapter test suite (tests/adapters/codex-cli/), install/update smoke (tests/scripts/codex-cli-install.test.sh), and a four-platform parity gate that proves Codex changes do not regress Claude Code, Gemini CLI, or OpenCode - Docs: new codex-cli.md guide, codex-migration.md for users switching from other platforms, and README/getting-started/examples updated for four-platform positioning - Bash harness: .gitattributes added to enforce LF on .sh files; harness LF-normalized so tests/run.sh works on Windows checkouts - .gitignore: .planning/ added (internal GSD workflow artifacts) Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>
8.1 KiB
Real-World Usage Examples
Practical scenarios showing how the Crew works in daily life. Each example shows what you say and what happens behind the scenes.
Scenario 1: The Brain Dump
Situation: You just got out of a meeting and your head is full of scattered thoughts.
You say:
"Quick dump: Marco wants the API docs by Thursday, Lisa mentioned the budget is getting cut by 15%, I had an idea about using webhooks instead of polling for the notification system, and I need to book a dentist appointment"
What happens:
- Scribe detects 4 distinct items
- Creates 4 separate notes:
Task: API Docs for Marco.md(with deadline: Thursday, linked to[[Marco]])Note: Budget Cut 15 Percent.md(tagged with #budget, linked to[[Lisa]])Idea: Webhooks for Notifications.md(tagged with #architecture, #notifications)Task: Book Dentist Appointment.md(tagged with #personal, #health)
- All notes land in
00-Inbox/with proper frontmatter - Scribe reports: "Captured 4 notes: 2 tasks, 1 idea, 1 informational note. Shall I save them?"
Scenario 2: Evening Inbox Triage
Situation: End of the day. Your inbox has accumulated 12 notes.
You say:
"Triage my inbox"
What happens:
- The
/inbox-triageskill scans all 12 notes in00-Inbox/ - Reads each note's content and frontmatter
- Files them:
- 3 meeting notes go to
06-Meetings/2026/03/ - 2 project tasks go to
01-Projects/Rebrand/ - 1 person note goes to
05-People/ - 2 ideas go to
03-Resources/Ideas/ - 2 ambiguous notes are kept in inbox with questions for you
- 3 meeting notes go to
- Updates 3 MOCs that gained new entries
- Leaves a message for Connector to check new links
- Reports the full summary with a clear breakdown
Scenario 3: Meeting Transcription
Situation: You recorded a 45-minute sprint planning meeting and pasted the raw transcript.
You say:
"Transcribe this meeting, it was the Q2 sprint planning with Marco, Lisa, and Ahmed. We met today at 10am."
[pastes raw transcript]
What happens:
- The
/transcribeskill processes the raw text - Identifies speakers from context
- Generates:
- Executive summary (4 sentences)
- 8 key discussion points
- 3 decisions made (with who decided and conditions)
- Action items table (6 tasks, assigned to specific people, with deadlines)
- Detailed notes organized by topic
- 2 open questions for follow-up
- Creates wikilinks to all participants:
[[Marco]],[[Lisa]],[[Ahmed]] - Saves to
00-Inbox/as2026-03-21, Meeting, Q2 Sprint Planning.md - Leaves message for Postman: "Check if there were pre-meeting emails about Q2 planning"
Scenario 4: Email Triage
Situation: Monday morning. You want to know what's important in your email.
You say:
"Check my email for anything urgent"
What happens:
- The
/email-triageskill scans your Gmail inbox (last 48 hours) - Reads 34 emails
- Filters:
- 22 newsletters/promos/notifications are ignored
- 3 action requests are saved as notes with tasks
- 2 deadline reminders are saved with deadline tags
- 4 informational emails from key contacts are saved as reference notes
- 1 meeting invitation is saved and flagged for calendar sync
- 2 ambiguous emails are summarized for your decision
- All notes land in
00-Inbox/ - Creates wikilinks to people in
05-People/ - Reports: "34 emails scanned. 10 saved to vault (3 urgent, 2 deadlines, 4 info, 1 meeting). 22 filtered out. 2 need your input."
Scenario 7: Knowledge Discovery
Situation: You're writing a proposal and need context from your notes.
You say:
"What do I know about microservices architecture? I need to write a proposal."
What happens:
- Seeker searches your entire vault
- Finds 7 relevant notes across different areas:
- 2 meeting notes where architecture was discussed
- 1 resource note on microservices patterns
- 3 project notes that reference architectural decisions
- 1 archived note from a past project
- Synthesizes: "Based on your notes, here's what you've documented about microservices..."
- Provides a structured summary with source citations (
[[Meeting: Architecture Review]], etc.) - Identifies a gap: "Note: your vault has no notes on service mesh or container orchestration. You might want to add these to your proposal research."
Scenario 8: Weekly Review
Situation: Sunday morning. Time for your weekly vault check-up.
You say:
"Run the weekly review"
What happens:
- The
/vault-auditskill runs a multi-phase audit:- Structure scan: all folders intact
- Duplicate detection: found 1 near-duplicate
- Link integrity: 3 broken links fixed automatically, 2 orphan notes flagged
- Frontmatter audit: 5 notes missing tags (auto-fixed)
- MOC review: 2 MOCs updated with new entries
- Growth analytics: 23 notes added this week, vault health score: 94%
- Generates a health report saved to
Meta/health-reports/ - Reports: "Your vault is in great shape! 1 duplicate needs your decision, 2 orphan notes might need homes. Everything else is handled."
Scenario 10: Graph Intelligence
Situation: You want to understand how your knowledge connects.
You say:
"Analyze my vault graph and find missing connections"
What happens:
- Connector runs a full graph audit
- Reports:
- 247 notes, 412 links, average 1.7 links per note
- 18 orphan notes (7%, under the 10% target)
- 3 isolated clusters
- Top connected note: your MOC/Engineering with 23 links
- Discovers 12 suggested connections:
- 4 strong (should definitely be linked)
- 5 medium (probably useful)
- 3 serendipitous (unexpected but interesting)
- Presents each with explanation: "Your note on 'Team Communication Patterns' should link to 'Sprint Retrospective Q1', because both discuss the same communication breakdown from different angles."
Codex CLI session examples
These examples show how to start and use the Crew with Codex CLI specifically.
Install and launch
# Install Codex CLI
npm i -g @openai/codex@latest
# Install the Crew for Codex CLI
bash scripts/launchme.sh --platform codex-cli
# Launch Codex in your vault
codex -C /path/to/your-vault
Verify the layout before your first session
# Non-interactive discovery smoke — confirms agents, skills, and dispatcher are visible
codex exec -C /path/to/your-vault "List the project custom agents under .codex/agents, the repo skills under .agents/skills, and the dispatcher file used in this workspace."
# Check MCP server visibility
codex -C /path/to/your-vault mcp list
Using agents and skills inside Codex
Once inside the interactive codex session, the Crew works the same way as on other platforms — just talk naturally:
"Initialize my vault" → /onboarding skill starts
"Save this note: quick idea" → Scribe agent captures it
"Triage my inbox" → /inbox-triage skill runs
"Check my email" → /email-triage skill scans Gmail
"Weekly review" → /vault-audit skill audits vault
Update after a git pull
cd /path/to/your-vault/My-Brain-Is-Full-Crew
git pull
bash scripts/updateme.sh --platform codex-cli
For the full runtime smoke matrix covering all 8 agents, all 14 skills, bounded child-agent chaining, and MCP visibility, see docs/codex-cli.md.
Daily Workflow Cheat Sheet
| Time | What to say | Skill/Agent |
|---|---|---|
| Morning | "What's on my calendar today?" | Postman agent |
| Morning | "Check my email" | /email-triage skill |
| Any time | "Save this: [thought]" | Scribe agent |
| After meeting | "Transcribe this meeting: [paste]" | /transcribe skill |
| Evening | "Triage my inbox" | /inbox-triage skill |
| Weekly | "Weekly review" | /vault-audit skill |
| When lost | "Find [topic]" | Seeker agent |
| When stuck | "Find connections for [note]" | Connector agent |
These are just examples. The beauty of the Crew is that you don't need to memorize commands. Just talk naturally, and the right agent will pick it up.