agent-lab-opencode AGENTS.md

Project instructions for agent-lab-opencode, an experiment runner for testing AI agents with different prompts and models in isolated Docker environments. Docker containers are separated environments used to run software safely and consistently.

In plain words
What is it for?
Use them when developing or changing the agent-lab-opencode project. They guide module design, code navigation, experiment setup, and verification.
Why use it?
They give agents a fixed engineering process, including defining a module's contract before coding, maintaining a project map, and treating verification as part of the work.

Instructions file for CodexOpenCode

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add instructions/sks-other/agent-lab-opencode/agents-md
Clone the repo
git clone --depth 1 https://github.com/SkS-Other/agent-lab-opencode

Made for: Codex, OpenCode.

Per session 2,355 This file is loaded in full into every session.
When invoked 2,355 The same file — it is already loaded in full.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.02355 $0.02355
Opus 5 $0.01177 $0.01177
Sonnet 5 $0.00471 $0.00471
Haiku 4.5 $0.00235 $0.00235

Measured yesterday against content hash 083f9c28c827, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

agent-lab-opencode AGENTS.md scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

AGENTS.md · 213 lines

How it starts

The opening of the file, as written. The whole thing — 213 lines — stays where its author put it; the contents beside it link to each section on GitHub.

GRACE Framework - Project Engineering Protocol

Keywords

agent, lab, llm, orchestration, sandbox, prompt, tests, opencode, configuration, tools

Annotation

CLI-driven, OpenCode-centered experiment runner: vary system prompt / model / task prompt, run isolated in Docker ("vacuum"), capture and compare agent behavior. Agents are the primary callers; humans secondary.

Core Principles

1. Never Write Code Without a Contract

Before generating or editing any module, create or update its MODULE_CONTRACT with PURPOSE, SCOPE, INPUTS, and OUTPUTS. The contract is the source of truth. Code implements the contract, not the other way around.

2. Semantic Markup Is Load-Bearing Structure

Markers like // START_BLOCK_<NAME> and // END_BLOCK_<NAME> are navigation anchors, not documentation. They must be:

  • uniquely named
  • paired
  • proportionally sized so one block fits inside an LLM working window

3. Knowledge Graph Is Always Current

docs/knowledge-graph.xml is the project map. When you add a module, move a module, rename exports, or add dependencies, update the graph so future agents can navigate deterministically.

4. Verification Is a First-Class Artifact

Testing, traces, and log anchors are designed before large execution waves. docs/verification-plan.xml is part of the architecture, not an afterthought. Logs are evidence. Tests are executable contracts.

5. Top-Down Synthesis

Code generation follows: RequirementsAnalysis -> TechnologyStack -> DevelopmentPlan -> VerificationPlan -> Code + Tests

Never jump straight to code when requirements, architecture, or verification intent are still unclear.

6. Governed Autonomy

Agents have freedom in HOW to implement, but not in WHAT to build. Contracts, plans, graph references, and verification requirements define the allowed space.

Grep-First Navigation

Use shared docs and semantic anchors as the primary navigation surface. Prefer grep and exact-text lookup before broad prose reading.

Read the full file on GitHub · 213 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. yesterday First seen · 213 lines · 2,355 tokens per session scan A 083f9c28c827

Subscribe to this mod's changes

agent-lab-opencode AGENTS.md is an instructions file published in the GitHub repository SkS-Other/agent-lab-opencode (0 stars, last pushed 2mo ago), licensed MIT. It adds 2,355 tokens to every session, about $0.0118 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Related

Other instructions, from other repositories

fastapi-langgraph-agent-production-ready-template AGENTS.md

Instructions for wassim249/fastapi-langgraph-agent-production-ready-template, covering ai agent development guide, quick commands, project structure, project overview and quick reference: critical rules.

wassim249/fastapi-langgraph-agent-production-ready-template · 1,955 tokens

MultiGen CLAUDE.md

Instructions for lingyuanli/MultiGen, covering claude.md, architecture, development commands, full stack (docker) and api (local).

lingyuanli/MultiGen · 1,388 tokens

agent-harness-kit AGENTS.md

Instructions for enmanuelmag/agent-harness-kit, covering agents.md — @cardor/agent-harness-kit, project, health check (run before making codebase changes), harness data (source of truth) and mcp tools (preferred).

enmanuelmag/agent-harness-kit · 780 tokens

openharness AGENTS.md

Instructions for mifunedev/openharness, covering open harness — orchestrator, what open harness is, 1. agent work stays inside the sandbox, 2. coding-harness choice does not change the workspace and 3. remote and unattended operation are normal.

mifunedev/openharness · 2,014 tokens

docker-nixuser AGENTS.md

Instructions for grigio/docker-nixuser, covering development guide, ci publishing (critical), multi-platform ci builds, flake auto-update and docker image.

grigio/docker-nixuser · 1,391 tokens

selfhost-ai CLAUDE.md

Instructions for kossakovsky/selfhost-ai, covering claude.md, project overview, core architecture, key files and installation flow.

kossakovsky/selfhost-ai · 5,945 tokens