# henriquetell

> Open-source publisher. Listings imported from github.com/henriquetell — credited to the original author with their license.

- **Listings:** 2
- **Total installs:** 0
- **Profile:** https://agentstack.voostack.com/s/henriquetell
- **Website:** https://github.com/henriquetell

## Published listings

- [Fable Mode](https://agentstack.voostack.com/l/skill-henriquetell-claude-skills-fable-mode) — Skill · Free · security-reviewed — `agentstack add skill-henriquetell-claude-skills-fable-mode`
  Make Claude Opus 4.8 work with Claude Fable 5's operating discipline — spec-first execution, aggressive tool/subagent use, fresh-context self-verification, grounded progress claims, memory, and calibrated autonomy. Use when the session runs on Opus 4.8 (or any non-Fable model) and the task is substantial — multi-step coding, long-horizon agentic work, large refactors, audits, deep research, overn…
- [Sonnet Lean](https://agentstack.voostack.com/l/skill-henriquetell-claude-skills-sonnet-lean) — Skill · Free · security-reviewed — `agentstack add skill-henriquetell-claude-skills-sonnet-lean`
  Cut token spend whenever work runs on Claude Sonnet 5 — as a delegated subagent/worker or as the session model. Applies Anthropic's documented Sonnet 5 cost levers — effort down-mapping (Sonnet 5 medium ≈ Sonnet 4.6 high), strict-effort scoping, thinking-trigger steering, verbosity control, one-shot complete briefs, and output-shape constraints. Trigger when delegating tasks to Sonnet workers (Ag…

---
Seller on AgentStack — the marketplace for AI agent skills and MCP servers. Every listing is security-reviewed. Install any with `agentstack add <slug>`.