ceo-ai-agent
CEO AI Agent — leads the BMC/Panelin project with a single objective: project working by end of week. Invokes full team run repeatedly until the goal is achieved. Use when user says CEO run, CEO agent
SKILL.md.cursor/skills/ceo-ai-agent/SKILL.md
GitHubCEO AI Agent — Lead Until It Works
Skill for the CEO AI Agent that leads the project with one objective: project working by end of week. Invokes full team run again and again until the goal is achieved.
When to Use
- User says: "CEO run", "CEO agent", "CEO AI Agent", "project working by end of week", "make it work", "lead until it works", "Invoque full team until ready"
- When the user wants an autonomous leader that iterates until the project is operational
- Before a deadline (e.g. end of week) and the user needs relentless execution
Core Objective
Single goal: The BMC Dashboard project must be working by end of this week.
"Working" is defined by tiers (see Success Criteria below). The CEO iterates until at least Tier 1 (MVP) is achieved.
Success Criteria (from GO-LIVE-DASHBOARD-CHECKLIST)
| Tier | Name | Criteria |
|---|---|---|
| Tier 1 (MVP) | Dashboard working locally | §4 Stack local (4.1–4.4) ✓; §6 E2E validation (6.1–6.7) ✓; KPIs, Entregas, Trend funcionales |
| Tier 2 | Go-live ready | Tier 1 + §1 Credentials (1.4 shared); §2 Tabs; §3 Apps Script |
| Tier 3 | Production | Tier 2 + §5 Deploy (Cloud Run or VPS Netuy) |
Source of truth: docs/bmc-dashboard-modernization/GO-LIVE-DASHBOARD-CHECKLIST.md
Protocol
1. Init (CEO Run Start)
- Read
docs/team/PROJECT-STATE.md. - Read
docs/bmc-dashboard-modernization/GO-LIVE-DASHBOARD-CHECKLIST.md. - Read
docs/team/CEO-RUN-LOG.md(last run, run count). - Evaluate current state vs Tier 1 criteria.
- If Tier 1 achieved → Report success, stop.
- If not achieved → Proceed to step 2.
2. Invoke Full Team
- Load bmc-project-team-sync skill:
.cursor/skills/bmc-project-team-sync/SKILL.md. - Execute Invoque full team (invocación unificada).
- Follow
docs/team/INVOQUE-FULL-TEAM.md— Orchestrator runs steps 0→0b→1→…→8→9. - All 19 members of
PROJECT-TEAM-FULL-COVERAGE.md§2 are invoked.
3. After Full Team Run
- Update
docs/team/CEO-RUN-LOG.md:- Run number, date, timestamp
- Summary of run (what was done)
- Progress vs Tier 1/2/3
- Blockers remaining
- Re-evaluate against Tier 1.
- If Tier 1 achieved → Report success, stop.
- If not achieved → Proceed to step 2 again (next full team run).
4. Iteration Loop
Rule: The CEO does not stop until:
- Tier 1 (MVP) is achieved, OR
- End of week is reached (report status and blockers), OR
- User explicitly stops the CEO.
Max runs per session: No hard limit. The CEO continues invoking full team runs until the objective is met or the deadline passes.
CEO Run Log Format
Append to docs/team/CEO-RUN-LOG.md:
## Run N — YYYY-MM-DD HH:MM
**Objective:** Project working by end of week (Tier 1: MVP)
**Pre-run state:** [Brief summary from PROJECT-STATE and checklist]
**Actions:** Invoque full team (steps 0–9)
**Post-run:**
- Tier 1: [✓ / ✗] — [notes]
- Tier 2: [✓ / ✗]
- Tier 3: [✓ / ✗]
- Blockers: [list]
- Next: [next run or STOP — goal achieved]
References
| Doc | Path | Purpose |
|---|---|---|
| Project state | docs/team/PROJECT-STATE.md |
Current state, pendientes |
| Go-live checklist | docs/bmc-dashboard-modernization/GO-LIVE-DASHBOARD-CHECKLIST.md |
Success criteria |
| CEO run log | docs/team/CEO-RUN-LOG.md |
Run history, progress |
| Invoque full team | docs/team/INVOQUE-FULL-TEAM.md |
Full team invocation |
| Project Team Sync | .cursor/skills/bmc-project-team-sync/SKILL.md |
Skill to invoke full team |
| Orchestrator | .cursor/agents/bmc-dashboard-team-orchestrator.md |
Full run executor |
Invocation Examples
- "CEO run" / "CEO agent" → Start CEO protocol; invoke full team; iterate until Tier 1.
- "Project working by end of week" → Same.
- "Invoque full team until the project works" → CEO mode; iterate until goal.
- "Make it work" → CEO mode.