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/starsumi/sumi-docs/agents-mdgit clone --depth 1 https://github.com/starSumi/sumi-docsWrote 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/starsumi/sumi-docs/agents-md)<a href="https://agentmods.dev/instructions/starsumi/sumi-docs/agents-md"><img src="https://agentmods.dev/badge/instructions/starsumi/sumi-docs/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 | $0.00877 | $0.00877 |
| Opus 5 | $0.00439 | $0.00439 |
| Sonnet 5 | $0.00175 | $0.00175 |
| Haiku 4.5 | $0.00088 | $0.00088 |
Grade A, and why
sumi-docs 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.
This is a copy
86% identical to Sumi-Docs-MCP AGENTS.md — 14 lines differ, which has more behind it and is treated as the original. This page carries a canonical link to it rather than competing with it.
How it starts
The opening of the file, as written. The whole thing — 87 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Sumi Docs Workspace Instructions
Product contract
This workspace owns the Sumi Docs product. It contains a read-only MCP server, an Astro/Starlight website, and a versioned corpus contract. Source visibility does not authorize a tag, package, deployment, binary release, or archival of predecessor repositories.
Ownership
packages/mcp/: headless, stateless, read-only MCP data plane.apps/web/: trusted content rendering and corpus publication.packages/corpus-contract/: pure schemas, canonicalization, fixtures, and conformance helpers. No transport, filesystem, Astro, or parser behavior.docs/: reviewed product documentation and default self-hosted corpus..agents/skills/: reviewed, repository-scoped usage and contribution workflows. Cross-repository maintainer policy remains user-scoped..codex/,.mcp.json,.vscode/: thin project MCP adapters only.
Nested AGENTS.md files continue to govern package-specific security and
validation. The root contract wins for workspace topology and cross-package
commands.
Reconciliation model
- For any mutable orchestration resource, keep versioned desired
specseparate from controller-ownedstatus. Machine-readableconditionsbindobservedGenerationto thegenerationactually observed and record type, status, reason, and immutable evidence. Stale or incomplete evidence isUnknown, never success. - Automation follows an observe-diff-reconcile loop. Reconciliation is idempotent, cancellation-aware, and restartable. Each mutable resource has one logical writer protected by a lease and monotonically increasing fencing token. Promotion uses compare-and-set against the expected generation or revision; conflicts and unavailable preconditions fail closed.
apps/webowns staging, sealing, atomic promotion, and rollback of complete publication artifacts. Any future controller owns leases, events, retries, checkpoints, and cleanup outside the repository. Skills route intent and procedure only;packages/mcpremains a stateless, read-only data plane.- Preserve the last verified artifact when reconciliation fails, and treat rollback as an explicit, evidence-backed generation transition. Do not add a generic orchestration API, distributed state store, scheduler, or extension framework without a measured workload and an accepted decision.
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 · 87 lines · 877 tokens per session scan A 0c3f89af1c4d
sumi-docs AGENTS.md is an instructions file published in the GitHub repository starSumi/sumi-docs (1 stars, last pushed 3d ago), licensed MIT. It adds 877 tokens to every session, about $0.0044 per session on Opus 5. A static security scan graded it A with 0 findings. It is 86% identical to Sumi-Docs-MCP AGENTS.md, differing in 14 lines, and is treated as a copy.
Other instructions, from other repositories
Sumi-Docs-MCP AGENTS.md
AGENTS.md instructions for starSumi/Sumi-Docs-MCP, covering sumi docs workspace instructions, product contract, ownership, reconciliation model and invariants.
Sumi-Docs-MCP CLAUDE.md
Claude Code instructions for starSumi/Sumi-Docs-MCP, a project described as: Read-only MCP server and Astro documentation system for local and remote documentation corpora.
DocGraph AGENTS.md
Instructions for Detective-XH/DocGraph, covering agents.md — docgraph fit guide for llm agents, the fit decision, decisive signals — any one means: use docgraph, supporting signals — useful when several hold together and anti-signals — prefer your own grep/read.
docgraph CLAUDE.md
Claude Code instructions for carrilloapps/docgraph, a project described as: Universal documentation knowledge-graph MCP server with hybrid full-text + vector search.
ui.plan.ai AGENTS.md
AGENTS.md instructions for planailabs/ui.plan.ai, covering agent guide, toolchain, layout, commands and skills.
GPT-RAG release.instructions.md
Instructions for Azure/GPT-RAG, a project described as: Enterprise-grade accelerator for agentic RAG on Azure. Built on Microsoft Foundry with Foundry IQ as the default retrieval backend, Microsoft Agent Framework orchestration, Zero-Trust architecture and IaC.