Agent Harness Construction
Design better AI agents: action space design, tool definitions, observation formatting, error recovery contracts, and context budgeting for higher completion rates.
- What
- Design better AI agents: action space design, tool definitions, observation formatting, error recovery contracts, and context budgeting for higher completion rates.
- Cost
- Free
- Needs
- Use "Agent Harness Construction" with your Muse.
- Install
- Copy the installer prompt below into your Muse — your agent does the rest.
Curated by Skill Harbor: a skill for people who build AI agents and want them to actually finish the job. It treats agent quality as a design problem constrained by four things: action space quality, observation quality, recovery quality, and context budget quality. You get concrete rules: stable explicit tool names, schema-first narrow inputs, deterministic output shapes, micro-tools for high-risk operations (deploy, migration, permissions) and macro-tools only when round-trip overhead dominates; every tool response carrying status, a one-line summary, next actions, and artifact references; an error recovery contract for every error path (root-cause hint, safe retry instruction, explicit stop condition); and context budgeting that keeps the system prompt minimal, moves guidance into on-demand skills, prefers file references over inlined documents, and compacts at phase boundaries. By @affaan-m, listed here with credit to its creator. From the affaan-m/ECC repository (MIT). Honest caveats: design guidance, not a framework; better harnesses raise completion rates but cannot fix a weak model or a vague task. Skill Harbor never reviews the code, review it yourself before use.
Version:
Install
Copy the install package below, then paste it into MuseThe install prompt below already includes the vetting steps: your agent follows the community checklist before installing anything with executable code. Want more?
Use "Agent Harness Construction" with your Muse. Prerequisites: none to install. Pure design guidance; have your agent's tool definitions or architecture notes ready. 1. Open the skill: https://github.com/affaan-m/ECC/blob/main/skills/agent-harness-construction/SKILL.md and copy the full SKILL.md text. 2. Paste it into a chat with Muse and add: "Review my agent's action space and tool definitions against these rules: [paste or describe]." 3. Apply the fixes it suggests (narrow inputs, observation format, recovery contracts) one area at a time. Tip: start with the error recovery contract; it is where most agents silently fail. Safety: a skill is plain-text instructions; it runs nothing by itself. Review any tool permission changes before deploying an agent.
Saved to your recent installs. Find it anytime on /connect.
Questions
How do I install a build?
Every product page includes a copy-paste install prompt. Paste it into your Muse and it sets the build up for you — no manual configuration.
Where does my money go?
Straight to the seller. Skill Harbor never processes payments: checkout happens on the seller’s own page, usually Stripe.
What does the ✓ next to a creator’s name mean?
It means we confirmed the identity of the person behind the listing. It says nothing about the code itself — always check a build before installing it.