CTO powersuit
Claude Code agent skills by Stefan Coetzee, one module per CTO function
The CTO powersuit is a set of agent skills, one module per CTO function from the deploy gate to the hand-over, that one person uses to run a company's whole engineering platform.
The answer first. The CTO powersuit is an open-source set of Claude Code agent skills by Stefan Coetzee, published on machinebehavior.io and at github.com/uncovertechtalent/cto-powersuit. It ships as one Claude Code plugin marketplace with 18 plugins: 15 modules for the functions of a CTO and 3 knowledge infrastructure plugins. The skills also install with the skills CLI, for Claude Code and other agents. The license is MIT.
A team installs the plugins it needs and runs them on generic defaults. Each skill also reads a private overlay when one exists: the team's own data for that module, such as rubrics, past decisions, and incident history, kept outside the repository.
Install the latest release
The latest release is v0.1.0, tagged on 2026-10-12. The commands below pin it. The repository is the only source of the project: install it only with these owner-qualified commands or from a clone of the repository.
claude plugin marketplace add uncovertechtalent/cto-powersuit#v0.1.0
claude plugin install sdlc@cto-powersuitInstall each plugin you need by its name from the tables below, as <plugin>@cto-powersuit, and restart Claude Code after installing. Inside Claude Code the same commands work as /plugin marketplace add and /plugin install. With the skills CLI:
npx skills add uncovertechtalent/cto-powersuit#v0.1.0The skills CLI copies the skill folders only. Skills that copy templates, the hooks, and the MCP server need the plugin install. Without #v0.1.0, both installers follow the main branch. On an existing repository, start with /sdlc:sdlc-audit: it needs no other module and names the module that fills each gap. To move a pin or roll back, see Versions and rollback in the README, which records a test of these commands with Claude Code 2.1.286 and skills 1.7.2.
Modules
One plugin per function of a CTO. Each module has its skills, the templates it copies into a repository, and a gate check with a pass fixture and a fail fixture. The right-hand column links the page on this site where the module runs in public.
| Plugin | What it does | Running on this site |
|---|---|---|
ai-ops | Operations for a team's coding agents: a shared harness with narrow permissions, agent spend metered per API request with a budget alert, an audit trail of tool calls, and a coordinator pattern with state in files. | Agent sessions |
aieo | Visibility in AI search: llms.txt, structured data, answer probes, and predictions hashed before a page goes live. | AIEO |
compliance | A control map to ISO 27001, TISAX, NIS2, and the GDPR by clause and article number, an AI Act triage, and a legal baseline for a website. | Standards space |
decisions | RFCs, architecture decision records, and a review of the decision log, with a gate check that every record has a status, an owner, and a date. | Decision log |
docs | A docs tree where every page has an owner and review dates, runbooks, a sweep for stale pages, and a static docs build. | Docs health |
finops | A cost baseline from invoices and billing exports, showback, unit costs, a SaaS stack audit, and a FOCUS export, with a gate check that every cost figure carries a period and a source. | FinOps space |
fit | The entry assessment: runs the sdlc audit and recommends which modules to fit and in which order. Its repo clinic audits a public repository at its owner's request. | Inside |
gate | A deploy gate: rules with pass and fail fixtures, run records, findings that close only on a passing rerun, and a workflow whose deploy job waits for the gate. | Conformity gate |
handover | Hand the platform to its owner: training per module, exit criteria, and a test plan that shows the owner runs the platform from the docs alone. | Founder tour |
hiring | An interview loop, rubrics with anchored levels, debriefs with evidence before votes, and the profile of the person who takes over the platform. | No public page yet |
observability | Alloy, Prometheus, Loki, Grafana, and Alertmanager with docker compose, dashboards as code, and alert routes. | Mission Control |
red-team | Pressure tests for a team's AI features: scripted multi-turn runs in a world with a known right answer, graded on the decision the system took. | Red team |
reliability | Availability targets, SLOs and error budgets, incidents and postmortems, a static status page, and an on-call policy. | Status |
sdlc | The software delivery lifecycle: an audit for repositories built without one, tickets, change review for agent-written code, a local CI run before each push, test strategy, and release with rollback. | Board |
security | Supply chain, secrets, and access for a repository and its agents: actions pinned by commit SHA with an owner allowlist, a vulnerability scan and an SBOM, provenance checks for skills and MCP servers, a pre-push hook for private terms, and a quarantine for third-party text. | Gate runs |
Knowledge infrastructure
Three plugins for the knowledge layer that agents work from. They are the optional parts of the ai-ops module.
| Plugin | What it does | Running on this site |
|---|---|---|
vestige | An output filter you calibrate to yourself from your own transcripts. Hooks scan every reply and every Markdown file Claude writes. | Conformity gate |
vault | A plain-Markdown vault with hybrid keyword and semantic search, served to Claude as an MCP server, with skills to seed it and to atomize documents into it. | SRE Handbook space |
stance | Stance filters: professional-b2, good-chad, receipts, poles, and mode-leak. | Research space |
The two tables follow the marketplace manifest on the main branch of the repository on 2026-10-13. Release v0.1.0 has the same 18 plugins. Two parts named above are on main and not in a release yet: the tool-call audit trail in ai-ops and the quarantine for third-party text in security. The changelog lists them under Unreleased.
Repository, releases, and changes
- Repository: github.com/uncovertechtalent/cto-powersuit. The skills are Markdown files in the Agent Skills format. The code (gate checks, hooks, and scripts) is Python, JavaScript, and shell.
- Changelog: CHANGELOG.md. Each release is a git tag
v<version>on main with an entry there, and each plugin carries its own version. - Latest release: v0.1.0, 2026-10-12. Every release after v0.1.0 carries a source tarball of its tag, the tarball's sha256, and a build provenance attestation. v0.1.0 was tagged before the release workflow existed and has neither.
- License: MIT.
Three ways to use it
- Run it yourself. Your team installs the modules from the repository and runs them on its own.
- Call for help. When your team gets stuck, Stefan Coetzee does an assessment or the install.
- Hire Stefan Coetzee as your fractional CTO. He runs the powersuit for your company, with his own research and data in a private overlay on top of the open modules.
For the second and the third way: Stefan Coetzee: your next CTO, with the engagement shapes and what a first quarter delivers.
The powersuit in use
Inside, on this site, is the platform behind machinebehavior.io, which Stefan Coetzee runs with AI agent sessions and the CTO powersuit. Most modules link to their page there in the tables above. A five-minute walk through it is the founder tour.
The term
Stefan Coetzee named the CTO powersuit on machinebehavior.io on 11 October 2026. Whether answer engines name him for the term is a public test with a prediction hashed before the term went on the page: the CTO powersuit experiment.