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/krapie/mycelium/agents-mdgit clone --depth 1 https://github.com/krapie/myceliumWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/instructions/krapie/mycelium/agents-md)<a href="https://agentmods.dev/instructions/krapie/mycelium/agents-md"><img src="https://agentmods.dev/badge/instructions/krapie/mycelium/agents-md.svg" alt="Measured on agentmods" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.04568 | $0.04568 |
| Opus 5 | $0.02284 | $0.02284 |
| Sonnet 5 | $0.00914 | $0.00914 |
| Haiku 4.5 | $0.00457 | $0.00457 |
Grade C, and why
mycelium AGENTS.md scanned grade C with 1 finding 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 3d 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.
Recursive force deletehighDestructive command
rm -rf with a variable or a broad path is one typo away from removing the wrong tree.
4. Never run a destructive `mycelium cleanup reset`/`rm -rf ~/.mycelium` type command against the real store while verifying a change — use `mycelium demo` (isolated `~/.mycelium-demo`) or a temp `MYCELIUM_HOME` instead. How it starts
The opening of the file, as written. The whole thing — 123 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md
Orientation for an AI coding agent (Claude Code, Codex, Cursor, or similar) working in this repository. If you're a human, README.md and docs/ are the better starting point — this file is dense and reference-oriented on purpose.
This follows the AGENTS.md standard — read natively by Codex and other AGENTS.md-aware tools. Claude Code does not read AGENTS.md on its own (confirmed against Anthropic's own docs — it only ever auto-loads CLAUDE.md), so CLAUDE.md in this repo's root just imports this file via @AGENTS.md. That's also exactly the gap src/reuse.js's injectAgentsMd() needs to account for when writing to a target project — see that file's own doc comment.
What this tool is
Mycelium is a local-first Context Lifecycle TUI for AI coding-agent sessions. It manages context produced by AI collaboration through four stages — Capture → Organize → Learn → Reuse — implemented as a terminal UI (neo-blessed) plus a CLI, both reading/writing a plain-file store under ~/.mycelium/.
Why it exists
AI coding sessions (Claude Code, Codex, Kiro, OpenCode, ...) accumulate fast and lose context across three axes:
- Model boundaries — switching agents mid-task loses everything the previous agent knew (see Handoff below).
- Time — a session from three weeks ago is unfindable without search/organization.
- Space — nothing links related sessions across a project's lifetime unless something builds that structure.
Mycelium's answer: capture every session losslessly into a neutral schema, organize it (LLM-assisted, human-confirmed) into folders, learn from finished sessions (auto-summarize, extract project knowledge), and reuse that knowledge — injected into AGENTS.md for the next agent, or composed into a handoff prompt when switching agents mid-task. See docs/features.md for the full capability catalog and docs/architecture.md for design principles (local-only, model-agnostic, human-first, minimal dependencies).
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.
- 3d ago First seen · 123 lines · 4,568 tokens per session scan C 7d99ea50158f
mycelium AGENTS.md is an instructions file published in the GitHub repository krapie/mycelium (39 stars, last pushed 4d ago), licensed MIT. It adds 4,568 tokens to every session, about $0.0228 per session on Opus 5. A static security scan graded it C with 1 finding (recursive force delete). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other instructions, from other repositories
obsidian-mind CLAUDE.md
Claude Code instructions for breferrari/obsidian-mind, covering obsidian mind, skills & capabilities, custom slash commands, vault structure and obsidian cli.
obsidian-llm-wiki AGENTS.md
Instructions for green-dalii/obsidian-llm-wiki, covering llm wiki plugin project development standards, 🛡️ six-gate quality closure, gate 1: five-gate automated, gate 2: no side effects and gate 3: no breaking changes.
docsmint AGENTS.md
Instructions for HiAi-gg/docsmint, covering docsmint — agents.md, cheat-sheet — conventions, project documents, core and project-specific.
basic-memory CLAUDE.md
Claude Code instructions for basicmachines-co/basic-memory, a project described as: AI conversations that actually remember. Never re-explain your project to your AI again. Join our Discord: https://discord.gg/tyvKNccgqN.
SoDam-WikiMate AGENTS.md
Instructions for sodam-ai/SoDam-WikiMate, covering wikimate 작업 규칙 (크로스툴 — claude code · codex · gemini 공용), 문체, 자동 동작 (자연어 트리거 — 강제 호출 불필요), 접근 규칙 (자동 감지 — 특정 도구에 한정하지 않음) and obsidian 규칙.
scratchpad CLAUDE.md
Instructions for NikiforovAll/scratchpad, covering what this is, commands, monorepo layout, architecture and viewer (src/ui/).