Sign InOpen Brain
AI EngineerVideoSource Linked

The Dark Arts of Skill Engineering — Paul Bakaus, Renaissance Geek (Impeccable)

Reliable agent skills need more than prompts: use hooks as automatic guardrails, test across models, and compile harness-specific variants when instruction-following differs.

AI Engineer · Sep 21, 2026
Open Source Open MarkdownOpen JSON
Source Summary

Impeccable grew from a normalization prompt into a design skill with **two-agent critiques**, scripts and design hooks. Its hooks support Claude Code, Cursor, Codex and GitHub Copilot, while model- and harness-specific builds adjust instructions and tool syntax.

Practical Implication

Treat skills as executable harness components. Put recurring checks into **pre-tool or post-tool hooks**, personalize them with project rules, and test each target model instead of assuming one prompt behaves consistently everywhere.

Agent-Ready Context
Impeccable grew from a normalization prompt into a design skill with **two-agent critiques**, scripts and design hooks. Its hooks support Claude Code, Cursor, Codex and GitHub Copilot, while model- and harness-specific builds adjust instructions and tool syntax.

Treat skills as executable harness components. Put recurring checks into **pre-tool or post-tool hooks**, personalize them with project rules, and test each target model instead of assuming one prompt behaves consistently everywhere.

Pre-tool blocking may be necessary for weaker instruction followers, but it is more restrictive. Cross-harness packaging also remains fragmented: common installers do not preserve Impeccable's provider-specific directory layout, so it uses a custom CLI.
Connected Context · Feed7 Judgment

This sharpens skill engineering from prompt writing into software maintenance: recurring checks can live in hooks, critiques can use multiple agents, and each model or harness may need its own instructions and tool syntax. Portability therefore requires target-specific testing and packaging; stricter pre-tool enforcement may improve compliance but also constrains execution.

On the Maintenance and Co-evolution of Agent Plugins: An Empirical Study of Claude Code Plugin MarketplacesThe marketplace study provides ecosystem-scale support for treating skills as evolving software whose instructions and scripts require coordinated maintenance, matching Impeccable’s hooks, scripts, and provider-specific builds.WorldFlowAI/everything-claude-codeThe harness kit demonstrates the same combination of skills, hooks, subagents, rules, and evals, while its context warning reinforces selective adoption rather than enabling every component.Agents Are Where Microservices Were in 2015 — Roberto Milev & Uday Kanagala, NavanNavan’s independently testable skills and tool-boundary guardrails reinforce Impeccable’s use of hooks and per-target testing, while adding persistence, traces, and trajectory tests as production requirements.Coding Agents Don't Scale Themselves. Neither Do Your Teams. — Patrick Debois, TesslImpeccable exposes the provider-specific maintenance burden behind shared skills; the team-systems view makes curation, evaluation, and guardrail ownership an organizational consequence of that burden.
Context Map
agentcoding#skills#harness-engineering#subagents
Uncertainty
Pre-tool blocking may be necessary for weaker instruction followers, but it is more restrictive. Cross-harness packaging also remains fragmented: common installers do not preserve Impeccable's provider-specific directory layout, so it uses a custom CLI.