foundry AGENTS.md

Guidance for defining and coordinating subagents in Claude plugins, including their roles, inputs, outputs, dependencies, and execution order.

In plain words
What is it for?
Designing sequential or parallel agent workflows, writing subagent definitions, connecting skills, and adding validation or transformation stages.
Why use it?
It helps decide when a task needs several specialized agents and how those agents should pass work between one another.

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/agentic-insights/foundry/agents-md
Clone the repo
git clone --depth 1 https://github.com/Agentic-Insights/foundry

Made for: Codex, OpenCode.

Per session 2,718 This file is loaded in full into every session.
When invoked 2,718 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 $0.02718 $0.02718
Opus 5 $0.01359 $0.01359
Sonnet 5 $0.00544 $0.00544
Haiku 4.5 $0.00272 $0.00272

Measured 2d ago against content hash 942e92859b07, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

foundry 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 2d ago.

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 · 369 lines

How it starts

The opening of the file, as written. The whole thing — 369 lines — stays where its author put it; the contents beside it link to each section on GitHub.

AGENTS.md - Subagent Patterns & Orchestration

Guidance for designing subagents in Claude plugins. Subagents extend agent capabilities beyond skills—they orchestrate, validate, and transform.

When to Use Subagents

Use skills - Single focused task, atomic operation, reusable instruction set Use subagents - Multi-step orchestration, parallel processing, conditional logic, error recovery, calling other skills in sequence

Subagent Definition

Subagents are defined in agents/<name>.md with YAML frontmatter:

---
name: subagent-name                          # Must be unique in plugin
description: What this subagent does         # Shown in agent discovery
capability: orchestration|validation|transform|review  # Category
input: { type: "description", schema: "..." }         # What it accepts
output: { type: "description", schema: "..." }        # What it produces
calls: [skill-name, other-subagent-name]             # Dependencies
parallelizable: true|false                           # Can run instances in parallel
---

Core Patterns

1. Sequential Orchestrator

Subagent calls skills in order, passing output as input to next:

INPUT
  ↓
[Call Skill A] → Output A
  ↓
[Call Skill B] ← Input: Output A
  ↓
[Call Skill C] ← Input: Output B
  ↓
OUTPUT

Example: Database migration orchestrator

  • Skill 1: Validate schema
  • Skill 2: Generate migration script
  • Skill 3: Create backup

2. Parallel Processor

Subagent spawns multiple instances of same skill with different inputs:

INPUT: [item1, item2, item3]
  ↓
[Skill instance 1] → Output 1   |
[Skill instance 2] → Output 2   | (parallel)
[Skill instance 3] → Output 3   |
  ↓
Aggregate results
  ↓
OUTPUT: [result1, result2, result3]

Example: Skill reviewer (validate N skills in parallel)

  • Input: List of skill directories
  • Each instance: Validate one skill
  • Aggregate: Combined report

3. Conditional Router

Subagent examines input, decides which skill path to take:

Read the full file on GitHub · 369 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. 2d ago First seen · 369 lines · 2,718 tokens per session scan A 942e92859b07

Subscribe to this mod's changes

foundry AGENTS.md is an instructions file published in the GitHub repository Agentic-Insights/foundry (5 stars, last pushed 2mo ago), licensed Apache-2.0. It adds 2,718 tokens to every session, about $0.0136 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.