Living bibliography behind the plugin. Bias: agent harnesses, loops, multi-agent orchestration, and production prompting — not toy chatbot demos.
- Prompting Claude Opus 5 — verbosity, narration, scope, subagent spawn, self-correction, thinking-disabled artifacts
- Claude prompting best practices — cross-model Claude techniques
- Effort parameter — cost/quality control without disabling thinking
- Designing Your Own AI Harness — loops, compaction, subagent isolation, unattended safety
- How to Build a Coding Agent Harness: 6 Core Layers — provider, loop, tools, context, sandbox, permissions
- Prompt, Harness, and Loop Engineering — simple composable patterns; think/act/observe/verify
- Harness Engineering for AI Coding Agents — constraint harnesses, feedback loops, PEV, quality gates
- From Prompts to Harnesses — Prompt → Context → Harness shift; Fowler cybernetic governor framing
- Code as Agent Harness (guide) — arXiv 2605.18747 synthesis; perceive/plan/act; hooks
- LLM-as-Code / Agentic Programming — code-driven workflows; LLM as adaptive node, not orchestrator of everything
- Every Agent Has a Harness (deepset) — harness sovereignty, hooks, budgets, model independence
- Swarm management in agent harnesses (Arize) — lineage, steer/kill, depth caps, cleanup
- OpenAI: Orchestration and handoffs — handoffs vs agents-as-tools
- OpenAI Agents SDK: multi-agent — manager patterns, code-orchestrated parallelism
- Anthropic writer-verifier patterns (see Opus 5 multi-agent coordination notes)
- AGENTS.md — cross-tool repo instructions (30+ agents)
- Agent Skills (Anthropic) — SKILL.md progressive disclosure; open standard
- Cursor plugins reference — rules, skills, agents, commands, hooks
- Claude Code plugins —
.claude-plugin/plugin.jsonlayout
| Resource idea | GODMODE surface |
|---|---|
| Opus 5 spawn/verbosity/scope | skills/opus-prompting, rules/delegation-law.mdc |
| Six-layer harness | skills/harness-design, agents/harness-smith.md |
| PEV / external verify | skills/pev-cycle, skills/ship-gate, agents/verifier.md |
| Swarm lineage + caps | skills/subagent-swarm, agents/swarm-captain.md |
| Council / independent lenses | skills/warroom-council, /warroom |
| Handoffs vs tools | docs/architecture.md, adapters/openai-agents-sdk.md |
| AGENTS.md + SKILL.md | root AGENTS.md, skills/*/SKILL.md |