Pack 307 · AI · 16 items
Claude Code & Agent Harness Pack
13 prompts + 3 skills for coding agent harness builders (Claude Code, Codex) — real jobs from the week, with the steps, rules and output an expert would ask for. 7 written for this work, 9 more from the library.
€4one-time · 16 prompts & skills · ≈ €0.25 each
- Instant download
- Yours to keep
- Works in ChatGPT, Claude, Gemini
What's inside
Made for coding agent harness builders (Claude Code, Codex)
6 prompts + 1 skill- PromptCLAUDE.md Project Memory File From a Real Repo
Turns a repo tour into a short, high-signal CLAUDE.md (or AGENTS.md) that tells a coding agent how to build, test, deploy and avoid the traps — without bloating every session.
- PromptAgent Skill Folder Design — SKILL.md That Loads at the Right Moment
Designs one reusable agent skill: the description that makes the agent pick it at the right time, the SKILL.md body, and the scripts or references it should carry.
- PromptAgent Hooks Plan — Guardrails That Run Every Time
Decides which rules should become hooks (always run) versus instructions (agent decides), and drafts each hook: the event, the check, the message the agent sees.
- PromptSubagent Roster — Who Does What in a Multi-Agent Coding Setup
Designs a small set of specialised subagents (reviewer, tester, researcher…) with clear briefs, tool limits and hand-offs, so parallel work stays cheap and doesn't collide.
- PromptMCP Server Tool Design Review — Tools an Agent Can Actually Use
Reviews or designs the tools of an MCP server: names, descriptions, parameters and error messages written for a model, with fewer, sharper tools and safe defaults.
- PromptCoding Agent Eval Suite From Your Own Past Tasks
Builds a small, honest evaluation set from your real tickets so you can tell whether a new prompt, skill, model or harness change makes the agent better.
- SkillCoding Agent Harness Architect
A standing architect for your coding-agent setup: decides what goes in the project memory file, which skills, hooks, subagents and MCP tools you need, and keeps the harness small and measurable.
Bonus prompts & skills
7 prompts + 2 skills · free with the packFrom the wider PromptCorrectly library, picked for the same work. They come in the same .zip and the start file tells your AI when to use each one.
- SkillAgent Skill Author
Writes an installable Agent Skill (agentskills.io / Claude Code / Codex CLI) from a messy capability request. Use when a paste-prompt should become a standing specialist.
Library · In a pack - SkillAI Builder Eval Harness
Designs a small, honest eval harness for a prompt, skill, or agent so the builder stops shipping on vibes. Use before calling anything 'production'.
Library · In a pack - PromptAutonomous Agent System Prompt Architect
Designs a complete, production-grade system prompt for an autonomous agent including persona, tool contracts, guardrails, and stop conditions.
Library · In a pack - PromptCode-Writing Agent Plan-Then-Execute Protocol
Governs a coding agent to explore, plan, implement in small verifiable steps, and self-test before declaring done.
Library · In a pack - PromptSelf-Critique Verification Gate For Agent Output
Adds a rigorous self-review pass where the agent critiques its own draft action against a checklist before committing it.
Library · In a pack - PromptAgent Memory Architecture And Summarization Policy
Designs short-term, long-term, and episodic memory plus a summarization policy to keep an agent coherent across long sessions.
Library · In a pack - PromptAgent Prompt-Injection Red Team Generator
Generates a structured battery of prompt-injection and jailbreak attacks to stress-test an agent before deployment.
Library · In a pack - PromptCost And Token Budget Optimizer For Agent Loops
Analyzes an agent workflow and proposes concrete changes to cut token cost and step count without losing task quality.
Library · In a pack - PromptHuman-In-The-Loop Approval Workflow Specifier
Specifies where an autonomous agent must pause for human approval, what to present, and how to resume after a decision.
Library · In a pack
What you download
Everything comes in one pack (.zip) you keep. Start with 00-GIVE-THIS-TO-YOUR-AI.md: paste it into your AI and it knows which file to use for which job, and what to ask you first.
More for Developers · Productivity
agent-harness-builder/ ├─ 00-GIVE-THIS-TO-YOUR-AI.md ├─ START-HERE.md ├─ TIPS-AND-FOLLOW-UPS.md ├─ AI-CHEAT-SHEET.md ├─ prompts/ │ ├─ 01-claude-md-project-memory-file-from-a-repo.md │ ├─ 02-agent-skill-folder-design-skill-md.md │ ├─ 03-agent-hooks-guardrails-plan.md │ ├─ 04-subagent-roster-for-a-codebase.md │ ├─ 05-mcp-server-tool-design-review.md │ ├─ 06-coding-agent-eval-suite-from-real-tasks.md │ ├─ 07-autonomous-agent-system-prompt-architect.md │ ├─ 08-code-writing-agent-plan-then-execute-protocol.md │ ├─ 09-self-critique-verification-gate-for-agent-output.md │ ├─ 10-agent-memory-architecture-and-summarization-policy.md │ ├─ 11-agent-prompt-injection-red-team-generator.md │ ├─ 12-cost-and-token-budget-optimizer-for-agent-loops.md │ └─ 13-human-in-the-loop-approval-workflow-specifier.md └─ skills/ ├─ coding-agent-harness-architect/SKILL.md ├─ agent-skill-author/SKILL.md └─ ai-builder-eval-harness/SKILL.md
A look inside
ROLE: You are a staff engineer who sets up coding agents for teams. You know that the project memory file is read at the start of every session, so every line costs attention, and a vague line is worse than none. You write the file a new senior hire …
Questions
What do I get?
A .zip with 13 prompts + 3 skills: 6 prompts + 1 skill written for coding agent harness builders (Claude Code, Codex), plus 7 prompts + 2 skills from the wider library that fit the same work. Each prompt is a ready .md file, each skill a SKILL.md folder, plus 00-GIVE-THIS-TO-YOUR-AI.md (a start file your AI reads first), START-HERE.md, TIPS-AND-FOLLOW-UPS.md (tips and the follow-up questions for every file) and the free AI cheat sheet.
How do I use it?
Open ChatGPT, Claude or Gemini and paste 00-GIVE-THIS-TO-YOUR-AI.md, or attach the folder. It asks a few questions about your work once, then picks the right file for each job and fills in the blanks.
Is it a subscription?
No. €4 once, and the files are yours to keep. Nothing renews.
How do I download it?
Right after checkout, on the confirmation page. Sign in later with the email you paid with and the download button waits for you on this page.
Why this price?
A pack is priced by the items written for its trade: €0.50 a prompt and €1 a skill. The 9 bonus items come free on top. This one is €4 for 16 prompts and skills, about €0.25 an item.