PHIacademy

AI Engineering Claude C1 Harness Engineering Project — AI skill for Claude Code

AI community

How to install AI Engineering Claude C1 Harness Engineering Project

This entry records only its repository, not the path inside it, so there is no exact command to give. Open PHIacademy/ai-engineering-claude-c1-harness-engineering-project and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What AI Engineering Claude C1 Harness Engineering Project does

Four Claude Code systems from the Harness Engineering capstone: a stop_reason-driven agentic loop with tool design, a token-budget context strategy for long conversations, a Claude Code monorepo config (CLAUDE.md, path-scoped rules, commands, skills), and multi-shift Layer 3 orchestration with tiered state, crash recovery, and session forking.

Alternatives in AI

Debate — AI Debate Hub - Structured four-way debates between Claude, Sonnet, Gemini, and Codex 2.8k ★
Agor — Orchestrate Claude Code, Codex, and Gemini sessions on a multiplayer canvas 1.2k ★
Cwc Long Running Agents — Claude Code's built-in /goal command gives you a generator/evaluator loop out of the box: set a completion con 668 ★

README

Project: Harness Engineering with Claude and Claude Code

Submission for the Harness Engineering project: four completed reference systems from the course, each built, run, and verified end-to-end, plus an evidence-grounded reflection brief.

The four systems

#	System	Run with	Tests passing
1	Insurance Claims Intake Agent — a `stop_reason`-driven agentic loop with structured tool use	`python -m claims_intake.run --all`	29
2	Retail Support Context Strategy — reduces token load while preserving answerability	`python -m retail_context.run --all`	30
3	E-Commerce Team Claude Code Config — CLAUDE.md hierarchy, path-scoped rules, commands, skills	`python -m ecommerce_team_config .`	35
4	Multi-Shift Quality Monitoring — Layer 3 orchestration with tiered state, crash recovery, forking	`python -m shift_monitor run-shift ...`	33

Repository structure

`README.md` — you are here
`reflection-brief.md` — evidence-grounded reflection brief
system-1-claims-intake/
- `summary.md` — claim outcomes table
- `system1-tests.png` — pytest run (29 passed)
- `traces_claim_02_stolen_bike.jsonl` — per-claim stop_reason trace
system-2-retail-context/
- `budget.json` — token accounting / reduction %
- `context.md` — assembled context artifact
- `case_facts_call.json` — structured case-facts extraction call
- `eval.jsonl` — evaluation results
- `eval_control.jsonl` — control (facts block stripped)
- [system2-tests.png](./system-2-retail-context/system2-tests