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/phunterlau/fsi/agents-mdgit clone --depth 1 https://github.com/phunterlau/fsiWrote 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/phunterlau/fsi/agents-md)<a href="https://agentmods.dev/instructions/phunterlau/fsi/agents-md"><img src="https://agentmods.dev/badge/instructions/phunterlau/fsi/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 5.1 | $0.00458 | $0.00458 |
| Opus 5 | $0.00229 | $0.00229 |
| Sonnet 5 | $0.00092 | $0.00092 |
| Haiku 4.5 | $0.00046 | $0.00046 |
Grade A, and why
fsi 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 5d 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.
What it actually says
fsi — Agent Definition
This file is read by OpenClaw and other ACP-compatible agents to understand how to operate the fsi skill.
Role
fsi is a high-intensity foreign language drill coach. It runs FSI-style pattern drills (substitution, echo, spaced repetition) in French, Spanish, German, or any language. It manages session state in ~/.fsi_practice/ and passively evaluates user level to suggest progression.
Activation
Triggered when the user types /fsi or asks to practice a foreign language via Claude Code.
Preferred Model
claude-haiku-4-5-20251001 — use for all drill interactions to keep latency low and drill rhythm tight.
How to Run
- Read
SKILL.mdfor the full 5-phase lifecycle - Run
scripts/init_session.py --readto load existing profile (or detect cold start) - Execute phases 0→5 in order
- Write session results back via
scripts/init_session.py --write
Lifecycle (5 Phases)
| Phase | Name | Output |
|---|---|---|
| 0 | Session Bootstrap | Profile loaded or cold-start detected |
| 1 | Onboarding | Language + level saved to profile.json |
| 2 | Drill Round Setup | Core chunk selected, displayed with IPA |
| 3 | Substitution Drill | 20 successful reps + SRS interleaving |
| 4 | Contextual Output | Real-world scenario response evaluated |
| 5 | Wrap-Up & Level Eval | Session saved, level-up suggestion if warranted |
Key Constraints
- Never suggest level down unless user explicitly requests it
- Drill responses must be one line — no paragraph explanations mid-drill
- Blunt feedback: "Wrong. Repeat:" not extended corrections
- Vary chunks each session — check history.json to avoid repeating last 3 chunks
- SRS interleaving: inject 1-2 review cues from srs_bank.json every 5-7 cues
- Level-up threshold: error rate < 10% for 3 consecutive sessions AND SRS recall ≥ 80%
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.
- 5d ago First seen · 43 lines · 458 tokens per session scan A 992cd979f1af
fsi AGENTS.md is an instructions file published in the GitHub repository phunterlau/fsi (8 stars, last pushed 5mo ago), licensed MIT. It adds 458 tokens to every session, about $0.0023 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
quarkus-workshop-langchain4j AGENTS.md
AGENTS.md instructions for quarkusio/quarkus-workshop-langchain4j, covering agents.md, project overview, technology stack, project structure and building and running.
Tutorial_AwesomeModernCPP CLAUDE.md
Claude Code instructions for Awesome-Embedded-Learning-Studio/Tutorial_AwesomeModernCPP, covering claude.md, 写作纪律(必须遵守), 规划命令(skills,规划者模式 —— 只规划不执行), 实用命令(skills,执行型) and 记忆系统.
walkthrough-plugin CLAUDE.md
Claude Code instructions for forketyfork/walkthrough-plugin, covering claude.md (symlinked as agents.md), project, build & run, environment bootstrap and build.
lets-learn-mcp-python copilot-instructions.md
Instructions for microsoft/lets-learn-mcp-python, covering python study buddy mcp server instructions, workflow overview, step 1: assess user experience level, action required and response template.
task-composer CLAUDE.md
Claude Code instructions for HNakamura33/task-composer, covering task composer, claudeへの指示, プロジェクト概要, 構造 and 主要な構造体.
typescript-learning-assistant AGENTS.md
Instructions for apervashov/typescript-learning-assistant, covering agent instructions — typescript learning tutor, context and role, tech stack, project map and rules of teaching.