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/mirasoth/soothe/agents-mdgit clone --depth 1 https://github.com/mirasoth/sootheWhat 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 | $0.05108 | $0.05108 |
| Opus 5 | $0.02554 | $0.02554 |
| Sonnet 5 | $0.01022 | $0.01022 |
| Haiku 4.5 | $0.00511 | $0.00511 |
Grade A, and why
soothe 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.
How it starts
The opening of the file, as written. The whole thing — 297 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Soothe Development Guide
Binding conduct for all agents and human contributors working in this repository. Compliance is mandatory; deviations require explicit operator approval and a recorded IG.
What Soothe is — a goal-driven orchestration framework for 24/7 autonomous agents. It extends deepagents with durable planning, reentrant loop state, and remote agent interop across a one-way monorepo dependency DAG.
What this guide governs — code placement and package boundaries (§7b), persistence backend selection (§10), verification and release gates (§5, §14), documentation and docstring standards (§12, §17), and commit/release attribution hygiene (§11, §13).
How to use it — read §1–§17 before any non-trivial change. Substantial work requires an IG in docs/impl/; minor changes follow commit/PR context. Run ./scripts/verify_finally.sh before every commit. When in doubt, stop and ask.
⚠️ CRITICAL RULES
1. Implementation Guides (IGs)
- Substantial work: Create IG in
docs/impl/(IG-XXX-brief-title.md) to track scope - Minor changes: No verbose IG — use commit/PR context, or minimal stub only
2. Config Sync
The packaged templates under packages/soothe-daemon/src/soothe_daemon/setup/templates/ are the source of truth; config/templates/ holds symlinks to them.
When editing a packaged template, MUST also update config/develop/nano.yml (or soothe.yml) with matching structure.
3. Ecosystem First
Check langchain-core, langchain-community, deepagents before implementing anything:
- Tools:
BaseTool,@tool - Subagents:
SubAgent,CompiledSubAgent - MCP:
langchain-mcp-adapters - Memory:
deepagents.MemoryMiddleware
4. Test Location
Tests go in package directories: packages/<pkg>/tests/unit/ or tests/integration/ — NOT root tests/.
5. Verification Required
Run ./scripts/verify_finally.sh before ANY commit. Zero lint errors, all tests pass.
6. After Code Impl: Cleanse → Verify → Fix (MUST)
Before marking work done (commit, PR, or handoff), apply this sequence every time:
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.
- yesterday First seen · 297 lines · 5,108 tokens per session scan A 3bbd985d5912
soothe AGENTS.md is an instructions file published in the GitHub repository mirasoth/soothe (10 stars, last pushed yesterday), licensed MIT. It adds 5,108 tokens to every session, about $0.0255 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.
Other instructions, from other repositories
pydantic-ai-todo AGENTS.md
Instructions for vstorm-co/pydantic-ai-todo, covering agents.md, project overview, quick reference, architecture and key design: storage injection.
pydantic-ai-todo CLAUDE.md
Instructions for vstorm-co/pydantic-ai-todo, covering claude.md, what this project is, commands, structure and core pattern.
mantishack AGENTS.md
Instructions for deonmenezes/mantishack, covering rust/codex-rs, the codex-core crate, code review rules, crate api surface and model visible context.
langgraph AGENTS.md
Instructions for langchain-ai/langgraph, covering agents instructions, corridor security analysis, libraries and dependency map.
deepagents AGENTS.md
Instructions for langchain-ai/deepagents, covering global development guidelines for the deep agents monorepo, corridor security analysis, development workflow, suppressing ruff rules and pr conventions.
codex-autoresearch AGENTS.md
Instructions for TheGreenCedar/codex-autoresearch, covering agents.md, purpose and scope, canonical sources, stable product boundaries and source and package boundaries.