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/no-materials/hird/agents-mdgit clone --depth 1 https://github.com/no-materials/hirdWhat 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 | $0.01662 | $0.01662 |
| Opus 5 | $0.00831 | $0.00831 |
| Sonnet 5 | $0.00332 | $0.00332 |
| Haiku 4.5 | $0.00166 | $0.00166 |
Grade A, and why
hird 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 — 161 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md
This repository is maintained with help from AI coding agents (e.g. Codex/ChatGPT/Claude). This file defines how to make changes, what “done” means, and the project defaults we enforce.
North Star
- Keep core crates small, predictable, and long-lived.
- Prefer simple, explicit designs over clever ones.
- Avoid dependency creep; keep compile times and surface area under control.
- Optimize for long-term architecture over short-term compatibility; it’s OK to break callers to get the right core shape.
Non-negotiables (Definition of Done)
cargo fmtpasses.cargo clippypasses (-D warnings).- Public APIs are documented (types/functions; public fields/variants where it matters).
- Tests updated/added when behavior changes.
- Examples/benchmarks live in separate top-level workspace crates (no extra dev-deps in core crates).
- Never reference tickets, ADRs, plans, phases, or agent sessions in code, comments, docstrings, or commit messages. These are project-management artifacts that rot. Code should justify itself; context belongs in the commit message or PR description, not in the source tree.
Suggested commands:
cargo fmt --all
cargo clippy --workspace --all-targets --all-features -- -D warnings
cargo test --workspace --all-features
Before implementing a ticket
Before writing any code, read these files in order:
- The assigned issue (
bd show <id>). DECISIONS.md— locked architecture constraints you must not violate.phrasebook.md— canonical surface syntax conventions.- The parent epic ticket if one exists (for broader context and out-of-scope boundaries).
Rust workspace expectations
- MSRV is set in
Cargo.toml(rust-version, currently 1.97); keep it compatible. - Follow workspace lint policy (notably:
unsafe_code = "deny"andmissing_docs = "warn").
no_std policy (core crates)
- Default assumption for foundational crates:
#![no_std]whenever practical (useextern crate allocwhen needed). - Keep
stdbehind an explicitstdfeature flag when required. - Avoid
stdcollections inno_stdcrates; usehashbrown(andalloctypes) instead.
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 · 161 lines · 1,662 tokens per session scan A 76fb80e0805f
hird AGENTS.md is an instructions file published in the GitHub repository no-materials/hird (0 stars, last pushed 2d ago), licensed Apache-2.0. It adds 1,662 tokens to every session, about $0.0083 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
Edict AGENTS.md
Instructions for Sowiedu/Edict, covering agents.md, what is edict?, quick orientation, agent skills (deep-dive guides) and mcp server.
sounio CLAUDE.md
Instructions for Sounio-lang/sounio, covering claude.md, 1. calibration — read before any analysis, 2. project identity, 3. session bootstrap and 4. build & run.
sounio AGENTS.md
Instructions for Sounio-lang/sounio, covering agents.md, steps, purpose, recovery + remote-first bootstrap and repository identity.
sounio copilot-instructions.md
Instructions for Sounio-lang/sounio, covering sounio language instructions for github copilot, compiler, critical syntax differences from rust, effects system (required) and variables and types.
SLAC-MCP AGENTS.md
AGENTS.md instructions for DennisPrediger/SLAC-MCP, covering agents.md, project, purpose, tools and non-goals.
cli AGENTS.md
Instructions for googleworkspace/cli, covering agents.md, project overview, build & test, changesets and architecture.