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.
npx agentmods add instructions/webdevtodayjason/sub-agents/claude-mdgit clone --depth 1 https://github.com/webdevtodayjason/sub-agentsWhat 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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.01806 | $0.01806 |
| Opus 5 | $0.00903 | $0.00903 |
| Sonnet 5 | $0.00361 | $0.00361 |
| Haiku 4.5 | $0.00181 | $0.00181 |
Grade A, and why
sub-agents CLAUDE.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.
How it starts
The opening of the file, as written. The whole thing — 215 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Claude Sub-Agents Configuration
[!NOTE] This project is archived (June 2026) and no longer maintained. Claude Code now provides subagents, skills (which absorbed custom slash commands), hooks, and plugins+marketplaces natively, so this framework is no longer needed. See the README for migration pointers. The guidance below is preserved as historical context.
🚨 CRITICAL: Concurrent Execution for Maximum Performance
GOLDEN RULE: Execute ALL related operations concurrently in a single message for 2-3x performance improvement.
🔴 MANDATORY Concurrent Patterns:
- File Operations: Batch ALL reads/writes/edits in ONE message
- Agent Spawning: Launch ALL related agents simultaneously
- Command Execution: Run ALL bash commands together
- Memory Operations: Store/retrieve ALL data in parallel
⚡ Concurrent Execution Examples:
✅ CORRECT - Single Message with Multiple Operations:
// Execute everything concurrently
[Single Message]:
- Read("src/api/users.js")
- Read("src/api/products.js")
- Read("src/models/user.js")
- Task("api-developer: implement CRUD endpoints")
- Task("tdd-specialist: create unit tests")
- Bash("npm install express")
- Bash("npm run lint")
❌ WRONG - Sequential Operations:
// Never do this - wastes time!
Message 1: Read("src/api/users.js")
Message 2: Read("src/api/products.js")
Message 3: Task("api-developer: implement endpoints")
Message 4: Bash("npm install")
// This is 4x slower!
🤖 Automatic Agent Orchestration
You are an intelligent orchestrator. PROACTIVELY use specialized agents for every task:
Task Analysis Rules
When receiving ANY task, immediately:
- Identify Components: Break down the task into specialized areas
- Spawn Agents Concurrently: Use multiple agents in parallel
- Don't Ask Permission: Just use agents - that's what they're for!
Automatic Triggers
- Code Changes → Always spawn:
code-reviewer+test-runner - New Features → Always spawn:
project-planner+tdd-specialist+ relevant developer - Bug/Error → Always spawn:
debugger+test-runner - Documentation → Always spawn:
doc-writer+api-documenter - Deployment → Always spawn:
devops-engineer+security-scanner
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.
- 2d ago First seen · 215 lines · 1,806 tokens per session scan A 19677c4e19b2
sub-agents CLAUDE.md is an instructions file published in the GitHub repository webdevtodayjason/sub-agents (201 stars, last pushed 2mo ago), licensed MIT. It adds 1,806 tokens to every session, about $0.0090 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 2026-08-30.
Other instructions, from other repositories
claude-code-tresor CLAUDE.md
Instructions for alirezarezvani/claude-code-tresor, covering claude.md - development guide, 🎯 repository purpose, 🏗️ architecture, 🛠️ common development commands and building & testing.
agentic-workflow CLAUDE.md
Instructions for gtrabanco/agentic-workflow, covering claude.md, repository layout, working rules, authoring a skill and hand off, don't compose across a model/effort boundary.
Claude-PHP-SDK copilot-instructions.md
Instructions for claude-php/Claude-PHP-SDK, covering copilot instructions for claude php sdk, project goals, architecture & core components, directory structure (to be created) and key architectural patterns.
claude-capsule-kit CLAUDE.md
Instructions for arpitnath/claude-capsule-kit, covering capsule kit v4.0, rules, context system (capsule), wikilink layer and docs.
contexture universal.instructions.md
Instructions for AcKeskin/contexture, covering universal rules, autonomy default, canonical tool commands, git: read all pr comments and git: active pr summary.
contexture cpp.instructions.md
Instructions for AcKeskin/contexture, covering cpp rules, c++ concurrency, c++ const-correctness, c++ error paths and c++ gpu interop.