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/0xjunha/darc/agents-mdgit clone --depth 1 https://github.com/0xjunha/darcWhat 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 | $0.00704 | $0.00704 |
| Opus 5 | $0.00352 | $0.00352 |
| Sonnet 5 | $0.00141 | $0.00141 |
| Haiku 4.5 | $0.00070 | $0.00070 |
Grade A, and why
darc 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 2d 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
Darc
- Follow idiomatic Rust and standard library-first design.
- Prefer less code when behavior stays equally clear and correct.
- Avoid new dependencies unless clearly justified. When adding one, enable only the required Cargo features and disable default features unless they are necessary.
- DRY: Reuse existing abstractions before adding new ones.
- Keep code DRY and cohesive. If the same logic or data shape appears in more than one place, refactor it into a shared function, helper type, or module unless duplication is clearly simpler. Avoid redundant struct fields and duplicated derived state; store canonical data once and compute the rest from it, unless duplication is clearly justified by readability, performance, or API boundaries.
- Add a one-line doc comment to every major struct and function.
- Run Cargo fmt and clippy after every patch or refactor that touches Rust code.
- fmt:
cargo +nightly fmt - clippy:
cargo clippy --workspace --all-targets --all-features -- -D warnings -W clippy::all - Linux clippy: run
scripts/check-linux-clippy.shwhen a Rust change touchescfg(...), platform-specific code, or service/background-process code.
- fmt:
- Split crates by cohesive capability and clear dependency direction, not by file count or verb names alone.
- Prefer leaf capability crates under a thin orchestration/facade crate. Lower-level crates must not depend on higher-level workflow crates.
- Only create or keep a crate boundary when it gives one dominant reason to change, a small API surface, and reduced change-coupling.
- Extract shared models and helpers downward into lower-level crates instead of duplicating them or making parser/storage code depend on orchestration code.
- Do not use real or local data for test fixtures or examples; use synthetic placeholders instead (for example, never copy UUIDs from local session history).
- Use conventional commits.
- Format:
<type>(<scope>): <imperative summary> - Rules:
- Use lowercase for
typeandscope. typemust be one of:feat,fix,refactor,perf,docs,test,chore,build,ci.scopemust name the primary area affected, such as a crate, module, package, or feature. e.g.,cli,parser,storage,api(optional).summarymust be a short imperative phrase describing the change itself, not the intention behind it.- Do not end the summary with a period.
- Keep the subject line concise, ideally under 72 characters.
- One commit should represent one logical change. The message should describe that single change.
- Use lowercase for
- When preparing a PR or commit with user-visible behavior, CLI output, install/release behavior, docs users rely on,
or compatibility impact, update
CHANGELOG.mdunderUnreleased.- Use one concise user-facing bullet.
- Do not add changelog entries for pure refactors, tests, formatting, or internal-only cleanup.
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.
- 2d ago First seen · 38 lines · 704 tokens per session scan A 071a26fbe06b
darc AGENTS.md is an instructions file published in the GitHub repository 0xjunha/darc (11 stars, last pushed 2mo ago), licensed MIT. It adds 704 tokens to every session, about $0.0035 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
Zuiti-Desktop AGENTS.md
Instructions for ZuiTiti/Zuiti-Desktop, covering agents.md, 项目概览, 目录地图, 运行命令 and 编码约定.
langgraph AGENTS.md
AGENTS.md instructions for langchain-ai/langgraph, covering agents instructions, corridor security analysis, libraries and dependency map.
agentgateway copilot-instructions.md
Copilot instructions for agentgateway/agentgateway: Do not check for, speculate about, or report compilation errors during code review. Compilation diagnostics from review are frequently incorrect; rely on CI to detect and report compilation failures.
hatch3r CLAUDE.md
Instructions for hatch3r/hatch3r, covering hatch3r — development instructions, architecture, development commands, two-axis pillar framework (2.0.0) and orchestrator self-discipline (bypass protection).
aeon CLAUDE.md
Instructions for aeonfun/aeon, covering aeon, how aeon works, strategy, voice and soul file hierarchy (read in this order).
ken CLAUDE.md
Instructions for townsendmerino/ken, covering claude.md, what this is, repository ownership (read this first), commands and embedding parity & golden fixtures (now in aikit).