cali (renato caliari)

🛸 cali (renato caliari)

> 🇧🇷 more about, in brazilian portuguese: > • 💼 linkedin > • 📝 substack — texts, e-books, prompts

📋 sections

| | |
|---|---|
| 🧪 [what i do](#what-i-do) | ⚡ [ai · workflow & agent skills](#ai--workflow--agent-skills) |
| 🛠️ [what i build with](#️-what-i-build-with-lately) | 📦 [pinned projects](#-pinned-projects) |
| 📖 [free e-books](#-free-e-books-brazilian-portuguese) | 📝 [writing & thinking](#-writing--thinking) |
| 🔭 [explorations](#-current-explorations) | |

---

what i do

i explore how product teams and organizations can replace rigid frameworks with lighter, more adaptive ways of working. my path — software development, product management, hr leadership, self-management consulting, and education — gave me a front-row seat to the frictions that drain energy and multiply unnecessary complexity. today i support people at every level, from c-level to product and people teams, helping them design their own ways of working rather than selling ready-made models.

i help people and groups navigate complex contexts — in their relationship with themselves, in organizational settings, and in digital product creation.

my work moves across four fronts:

product discovery, job to be done & innovation

created the triple track agile approach. wrote *a e-book to building products with short learning cycles* and *strategy and innovation with jobs to be done*, in portuguese.
at timeproduto. i map product workflow stages, tools for each, and publish articles on product careers, design, dev, and ai impact — all in portuguese.
the stelow workflow and skills come from years inside real teams — consulting, leading, and experimenting across different organizations. they're not abstract theory.
→ github.com/renatocaliari · linkedin

#### ai agents & skills

🤖 30+ open skills for product discovery, coding, qualitative research, and workflow automation. an agent ecosystem.
→ github.com/calionauta/agent-sync-public

#### culture & organizational design

🔄 culture & organizational design — structural experiments, alternatives to traditional corporate rituals (performance reviews, 1:1s, 360s), decentralized decision-making through a complex systems lens. supporting c-level to team leads, teams, and individuals.
→ linkedin

#### participatory facilitation & narrative therapy

🌀 liberating structures, group dynamics, and original facilitation processes. training and certification in narrative therapy (in progress). i also build ai tools that support narrative practices (practice library, landscape compass, therapist training simulator
currently building tool for simulated conversation system with ai supervision for therapist training.
→ calionauta.com · linkedin

---

⚡ ai · workflow & agent skills

i create specialized ai agent skills for product discovery, qualitative research, and coding workflows. these are prompt-based tools that help agents execute specific tasks with precision.

featured: stelow — opinionated product workflow. stellar clarity. low friction.

stelow coordinates +20 skills (listed below). the design draws on 2025–2026 research — parallel orchestration (CAID), cross-session learning (Cat), visual review gates (Plannotator), AI-aware mutation testing. i also document 11 known limitations of what AI still gets wrong, even with this structure. it's designed to amplify human judgment, not replace it.

| layer | role | skills | count |
|-------|------|--------|:-----:|
| 🎛️ orchestrator | pipeline coordinator | stelow | 1 |
| 🧠 product strategies | upfront exploration | product-discovery · job-to-be-done · evolutionary-principles · opportunity-mapping · multi-method-market-analysis | 5 |
| ⚙️ workflow stages | pipeline steps | shape-up · plan-critique · codebase-critique · ux-critique · interface-alternatives · tech-planning · testing-ai-code · testing-execution · scope-executor · execution-critique | 10 |
| 📘 product tactics | on-demand reference | ads · pricing · promotions · trust-building · health · marketplace-playbook · business-models · open-source | 8 |
| 🔧 cross-cutting | shared conventions | code-standards | 1 |
| | | total | 25 |

other key skills

| skill | domain | description |
|-------|--------|-------------|
| cali-agents-md-generator | 📝 workflow | generate / maintain project AGENTS.md |
| cali-agents-md-validator | 📝 workflow | validate AGENTS.md against best practices |
| cali-coding-standards | 💻 coding | universal principles — KISS, DRY, LoB, SoC, Fail Fast + size limits & CI enforcement |
| cali-ops-deploy-github-tailscale | 🚀 ops | deploy to private servers via tailscale |
| cali-ops-docker-server-dashboard | 🚀 ops | server discovery & real-time docker dashboard |
| cali-ops-github-releases | 🚀 ops | github release automation |
| cali-coding-go-stack | 💻 coding | go web app conventions — datastar, templ, chi, sqlite |
| cali-coding-go-standards | 💻 coding | go project conventions and best practices |
| cali-ops-package-audit | 🚀 ops | supply chain security audit — socket.dev, trivy, osv-scanner, dotenvx |
| cali-questions-quality | 🎙️ research | evaluate interview scripts & qualitative research |
| cali-ops-server-security | 🚀 ops | server security audit and hardening |
| cali-skill-validator | 📝 workflow | validate agent skills against best practices |
| and more | … | … |

⚙️ the pipeline

> pi.dev + custom skills
>
> I set the stage upstream — questions, critiques, planning convos, detailed scopes. Then the AI runs inside the fences. I still sneak in when something feels off. The results surprise me more often than not.

---

🛠️ what i build with lately

🐹 go · ✨ templ · 🔀 chi · ⚡ datastar · 🗄️ sqlite

~ no javascript frameworks, no runtime deps, no huge node_modules — just type-safe go, server-side reactivity via SSE, and a single binary you can deploy anywhere.

---

📦 pinned projects

| project | description |
|---------|-------------|
| 🔄 agent-sync | agent configuration sync ecosystem |
| 🌐 agent-sync-public | public-facing agent sync & skill distribution |
| ⚡ datastar-lint | linter for datastar HTML attributes. language-agnostic. |
| 🐹 gogogo-fullstack-template | go web app template, batteries included. single binary, zero external services, llm-friendly |
| 🛡️ pi-leakguard | leak guard for pi.dev - blocks llm access to sensitive paths, egress dlp, taint tracking, and secret redaction. |
| 🔧 pi-tool-repair-layer | pi extension that fixes llm tool call errors |
| 🤖 stelow | opinionated agentic product workflow orchestrator for ai agents and humans. ai agents less like coding assistants, more like cross-functional product team |

---

📖 free e-books (brazilian portuguese)

| title | topic |
|-------|-------|
| 📘 guia prático para validar ideias | 8-step experiment framework — find audiences, define markets, prototype offers, and evolve into viable businesses without building full products first |
| 📘 estratégia e inovação com jobs to be done | comprehensive jtbd methodology in portuguese — job mapping, market segmentation, outcome-driven innovation adapted for real-world use |

---

📝 writing & thinking

i write at 🛸 espaço calionauta — reflections on philosophy, habits, narrative, culture, and life.

sections: 🧪 life lab · 🐙 tentaculus (culture, product, jtbd) · 🤖 ai prompts · 📜 quote collection

other places:

• 🌐 calionauta.com — narrative therapy & facilitation
• 💼 linkedin
• 🐦 @calionauta

---

🔭 current explorations

• narrative therapy & participatory narrative inquiry
• ai agents for therapeutic and facilitative contexts
• ai-human collaborative product workflows — designing systems where ai assists each stage while keeping human judgment at the center
• lighter organizational rituals and structures
• minimalism, strength training, whole food plant based
• life without goals — systems and practices instead

Collapse