Set up
Skills are installed by placing their folder where your agent loads skills.
The download comes from github.com/igor-elbert/ctobot itself, not from our server, so what you get is exactly what its author published.
Install instructions come from the README
README
synced Oct 21, 2026CTObot
An AI agent operating manual for the CTO role at small companies (≤ 50 engineers, Seed to Series C).
What this is
CTObot is a pair of structured markdown files that give an AI agent — or a human CTO — everything needed to reason and act as a Chief Technology Officer:
| File | Purpose |
|---|---|
AGENTS.md | The full operating manual: responsibilities, decision frameworks, thresholds, cadences, escalation rules, culture principles, and standard artifact templates |
SOUL.md | The CTO persona: identity, core truths, boundaries, and voice (soul.md/OpenClaw schema). The recurring-problem quick reference lives in CHEATSHEET.md |
Together they form a complete, opinionated CTO operating system — one you can load into an AI agent, hand to a new technical leader, or use as a personal reference.
Who it is for
- CTOs and VP Engs at Series A–C companies who want a structured operating model, not a generic leadership blog post
- Founders who are acting as de facto CTO and need a checklist-driven way to run the technology organisation
- AI practitioners who want to equip an agent with CTO-level context for strategy, risk, roadmap, and team questions
What it covers
Strategy & roadmap
- Three-horizon planning (H1/H2/H3) with a 70/20/10 portfolio split
- OKR mapping for all engineering initiatives
- Capacity conflict resolution framework
Engineering health
- DORA metrics with elite thresholds, acceptable floors, and action triggers
- Data infrastructure health metrics
- Weekly, monthly, quarterly, and annual review cadences
Technical debt & legacy systems
- Debt classification: deliberate, inadvertent, reckless
- Four-phase legacy remediation: knowledge extraction → containment → incremental replacement → business rules externalisation
Build vs. Buy
- Five-question decision test with 3-year TCO model
- Decision memo template with exit and rollback plan
Security
- Shift-left security checklist
- Zero-trust and least-privilege principles
- Rapid-growth warning triggers
Team & culture
- Hiring, performance management, and retention frameworks
- Commitment culture and a shared glossary
- Blameless post-mortems with a mandatory learning loop
Standard artifacts (all templates included)
- Architecture Decision Record (ADR)
- Build-vs-Buy decision memo
- Risk register
- Escalation memo
- Incident post-mortem
- System owner / co-pilot registry
- Weekly CTO review
- Monthly leadership tech summary
- First 30 / 60 / 90 day operating checklist
Skills package
CTObot also ships a reusable skills/ directory. These skills are meant to be copied into the dot-directory used by the tool you are running — for example .claude/skills/, .opencode/skills/, or ~/.hermes/skills/.
| Skill | What it does |
|---|---|
cto-challenge-extraction | Turns external CTO articles or interviews into a clean, deduplicated challenge list. |
cto-operating-doc-red-team | Red-teams AGENTS.md / SOUL.md against those challenges and flags real execution gaps. |
cto-red-team-scorecard | Scores each challenge by coverage, failure mode, missing artifact, severity, and pass/partial/fail. |
cto-challenge-stress-test | Stress-tests an agentic CTO against common challenge families and compares responses to the manual. |
proprietary-leak-scan | Pre-commit safety gate: flags company-specific content (org name, emails, NPI/PII, secrets) that has slipped into the public files. |
When you copy them into another tool’s dot-directory, keep the same layout: skills/<category>/<skill-name>/SKILL.md.
How to use it
With your favorite AI tool
Most tools (e.g. Claude Code, OpenCode, Hermes) are AGENTS.md, SOUL.md, and skills/ aware. Start your tool from the directory containing these files, or copy the skills/ tree into the tool’s dot-directory so the agent can discover them in its native location.
As a personal operating manual
Work through the First 90 Days Checklist on arrival. Use the templates in AGENTS.md §16 as your standard artifacts. Use CHEATSHEET.md to diagnose recurring problems and map each to the right AGENTS.md section.
As a team reference
Share with Engineering Managers and Staff Engineers as the canonical operating model. Anchor architecture reviews, incident post-mortems, and roadmap conversations to the frameworks here.
Status
Early / experimental. The frameworks are opinionated and have been assembled from cited public sources (linked inline in AGENTS.md). They reflect one coherent point of view — not universal truth. Use what works, adapt what does not, and share what you learn.
Last updated: April 2026
Contributing
This is a living document. Contributions that improve accuracy, fill gaps, or reflect hard-won experience are welcome.
To propose a change:
- Open an issue describing the problem or gap you have observed in practice
- Submit a pull request with the proposed edit to
AGENTS.md,SOUL.md, or the skills you suggest - In the PR description, state: what changed, why it is better, and whether it supersedes an existing section
Contribution standards:
- Changes to
AGENTS.mdmust cite a source or describe a concrete operational rationale - Changes to
SOUL.mdmust keep it a lean persona (identity, core truths, boundaries, voice); operating detail belongs inAGENTS.md - Suggested skills must be correctly formatted with
SKILL.mdand the explanation - Do not add frameworks that cannot be verified or acted on by a working CTO
The glossary in AGENTS.md §15.7 governs the meaning of terms used throughout both files. If a contribution introduces new terminology, add it there.
License
MIT — use freely, adapt openly, attribute where it matters.