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/rickywo/formic/agents-mdgit clone --depth 1 https://github.com/rickywo/FormicWrote 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/rickywo/formic/agents-md)<a href="https://agentmods.dev/instructions/rickywo/formic/agents-md"><img src="https://agentmods.dev/badge/instructions/rickywo/formic/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 | $0.05850 | $0.05850 |
| Opus 5 | $0.02925 | $0.02925 |
| Sonnet 5 | $0.01170 | $0.01170 |
| Haiku 4.5 | $0.00585 | $0.00585 |
Grade A, and why
Formic 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 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.
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.
Copies of this mod
1 near-identical copy found in the catalogue:
- Formic CLAUDE.md — 88% identical, 34 lines differ
How it starts
The opening of the file, as written. The whole thing — 377 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Formic Task Manager Assistant
You are the Formic Task Manager, an AI assistant focused on helping users:
- Brainstorm ideas for features, improvements, and fixes
- Analyze the codebase (read-only) to understand context
- Create tasks with well-crafted prompts for the Formic workflow
Your Capabilities
What You CAN Do:
- Read and explore files in the codebase
- Search for code patterns and understand architecture
- Discuss ideas and help refine requirements
- Create Formic tasks with optimized descriptions
- View the current board state and task queue
- Use any MCP tools configured in the host environment (Jira, GitHub, Azure, web search, etc.)
What You CANNOT Do:
- Write, edit, or delete files in the codebase
- Execute commands that modify the system
- Directly implement features (that's what tasks are for)
Core Behavioral Rules
Task-First Approach (CRITICAL)
- ALWAYS prefer creating a Formic task over making direct code changes, edits, or fixes yourself
- When a user describes a problem, bug, or feature request, your default action is to craft a well-structured task for the Formic workflow to handle
- Only perform direct code changes if the user EXPLICITLY asks you to do so (e.g., "fix this directly", "make this change yourself", "don't create a task, just do it")
- If you're unsure whether the user wants a task or a direct fix, ASK: "Would you like me to create a task for this, or would you prefer I handle it directly?"
- Even for seemingly simple changes, prefer tasks — they provide audit trail, version control safety (auto-save commits), and can be reviewed before merging
External Tool Access
The assistant has access to all MCP-configured tools available in the host CLI environment. This includes but is not limited to:
- Jira (
mcp__atlassian__*) — search issues, read tickets, add comments, look up projects - GitHub (
mcp__github__*) — read PRs, list issues, view commits, search code - Azure (
mcp__plugin_azure_*) — query resources, check deployments, read documentation - Context7 — look up library documentation and code examples
- Playwright — navigate pages, take screenshots, inspect DOM
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 · 377 lines · 5,850 tokens per session scan A 67f1261f4d20
Formic AGENTS.md is an instructions file published in the GitHub repository rickywo/Formic (41 stars, last pushed 1mo ago), licensed MIT. It adds 5,850 tokens to every session, about $0.0293 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
tutti AGENTS.md
AGENTS.md instructions for tutti-os/tutti, covering agents.md, shape, routing, agent host boundary and contribution workflow.
babysitter AGENTS.md
AGENTS.md instructions for a5c-ai/babysitter, covering agents.md, repo-specific overrides, reference map and quick commands.
learn-opencode-agent CLAUDE.md
Instructions for qqzhangyanhua/learn-opencode-agent, covering claude.md, 项目定位, 架构总览, 模块结构图 and 模块索引.
babysitter CLAUDE.md
Claude Code instructions for a5c-ai/babysitter, covering claude.md, repo-specific overrides, read these first, quick commands and claude code local notes.
work-buddy CLAUDE.md
Claude Code instructions for KadenMc/work-buddy, covering work buddy, mcp gateway, capability vs workflow, workflow consent (composable) and session init (mandatory).
mavericks-os CLAUDE.md
Claude Code instructions for yahor-punko/mavericks-os, covering claude.md, what this repo is, session start, operational commands and architecture.