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/edison7009/echobird/agents-mdgit clone --depth 1 https://github.com/edison7009/EchoBirdWhat 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 | $0.00786 | $0.00786 |
| Opus 5 | $0.00393 | $0.00393 |
| Sonnet 5 | $0.00157 | $0.00157 |
| Haiku 4.5 | $0.00079 | $0.00079 |
Grade A, and why
EchoBird 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 — 67 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md — EchoBird
Instructions for agents (ZCode / Claude Code / Codex) working in this repo. Read this before modifying code.
Project shape
Tauri desktop app: React + TypeScript + Vite frontend in src/, Rust backend
in src-tauri/. Two-tier tool/feature registries often ship in both a
bundled copy and a remote copy (e.g. src/data/<x>.json vs
docs/api/<...>/index.json) — when adding or editing an entry, update both
files unless told otherwise; they share one schema and the remote wins at
runtime. Prettier governs the frontend JSON/TS; rustfmt governs src-tauri/.
CI gates (must pass locally before pushing)
CI runs two jobs in parallel, each with this order — format runs first and fails the whole job before clippy/test even start, so a format slip hides whether the code is actually sound. Match the order locally:
Frontend (src/, repo root):
npm run typecheck # tsc --noEmit
npm run format:check # prettier --check (or `npm run format` to auto-fix)
npm run lint # eslint (or `npm run lint:fix`)
npm test # vitest
Rust (src-tauri/, or pass --manifest-path src-tauri/Cargo.toml):
cargo fmt --check # FORMAT FIRST. `cargo fmt` to auto-fix.
cargo clippy --all-targets --all-features -- -D warnings # CI denies warnings
cargo test
Run all of them, every push. A green cargo check/cargo test does NOT
imply a green CI — cargo fmt --check is a separate gate and the one most
often skipped. Same for the frontend: npm run format:check is not implied by
typecheck or lint.
Lesson (learned the hard way)
Commit 4774b1d5 passed cargo check + cargo clippy + cargo test locally
but failed CI because cargo fmt hadn't been run (a braced-import reorder and
a wrapped method chain). CI's cargo fmt --check failed in the first step,
so clippy/test never executed — the failure said nothing about correctness,
only about formatting. Don't repeat it: treat the format gate as equal to
clippy/test, not optional.
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 · 67 lines · 786 tokens per session scan A ec8c56b8eb69
EchoBird AGENTS.md is an instructions file published in the GitHub repository edison7009/EchoBird (3,146 stars, last pushed yesterday), licensed MIT. It adds 786 tokens to every session, about $0.0039 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
hotplex AGENTS.md
Instructions for hrygo/hotplex, covering hotplex 项目知识库, 约定与规范, 必须遵守, 反模式（禁止） and 代码编辑规则.
agentic-ai CLAUDE.md
Instructions for DjangoPeng/agentic-ai, covering claude.md, repository shape, common commands, financial automation and morning newspaper.
kaogong-study-tracker AGENTS.md
Instructions for KaguraNanaga/kaogong-study-tracker, covering 朱批录 agent instructions, when to trigger, core flow, local data and image capability.
kesha-voice-kit AGENTS.md
Instructions for drakulavich/kesha-voice-kit: Project rules, build commands, release process, and lessons learned live in CLAUDE.md — single source of truth for every coding agent (Claude Code, Cursor, Aider, Codex, etc.). When CLAUDE.md and any other doc disagree, CLAUDE.md wins.
causal-memory CLAUDE.md
Instructions for JingxuanC/causal-memory, covering causal memory mcp integration, causal memory integration, after acting on a decision and observing the result, when something fails unexpectedly and when a recorded lesson turns out to be wrong.
no-vibe CLAUDE.md
Instructions for rizukirr/no-vibe, covering claude.md, repo purpose, verification, architecture — parallel surfaces, one behavior and two data layers (easy to confuse).