Borrowing it
Nothing to install: this file belongs to jarvis-intelligence/jarvis. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/jarvis-intelligence/jarvis/main/AGENTS.mdgit clone --depth 1 https://github.com/jarvis-intelligence/jarvisWrote 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/jarvis-intelligence/jarvis/agents-md)<a href="https://agentmods.dev/instructions/jarvis-intelligence/jarvis/agents-md"><img src="https://agentmods.dev/badge/instructions/jarvis-intelligence/jarvis/agents-md/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/instructions/jarvis-intelligence/jarvis/agents-md"><img src="https://agentmods.dev/badge/instructions/jarvis-intelligence/jarvis/agents-md.svg" alt="Reviewed on agentmods" width="80" 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 | $0.05802 | $0.05802 |
| Opus 5.5 | $0.02321 | $0.02321 |
| Sonnet 5.5 | $0.01160 | $0.01160 |
| Haiku 4.5 | $0.00580 | $0.00580 |
Grade A, and why
jarvis AGENTS.md scanned grade A with 1 finding 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 1mo 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.
Runs shell commandslowCapability
Expected in a hook, worth knowing in a rule or an instructions file.
- **Unit tests mock every external boundary**: subprocess via monkeypatched `jarvis.index_cli._run`/`subprocess.run` (dispatch on `step.endswith(" index")`); HTTP via `httpx.MockTransport` mirroring Zoekt's `/api/search` How it starts
The opening of the file, as written. The whole thing — 123 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Repository Guidelines
Contributor guide for jarvis — a local-first code intelligence MCP server (SCIP symbol navigation + Zoekt lexical search + optional semantic vector search). PyPI distribution is jarvis-mcp; the import package, both CLIs, and the MCP server are all jarvis.
Project Overview
jarvis gives AI coding assistants code-intelligence tools over a local data dir (~/.jarvis, overridable via JARVIS_DATA_DIR). Two processes share that dir:
- Writer —
jarvisCLI (src/jarvis/index_cli.py): runs per-language SCIP indexers, converts to SQLite, builds Zoekt shards (+ optional LanceDB vectors), publishes atomically. - Reader —
jarvis-serverMCP stdio server (src/jarvis/server.py): 9 tools (documentSymbols, goToDefinition, findReferences, callHierarchy, typeHierarchy, getIndexStatus, searchCode, semanticSearch, blastRadius).
Single-tenant by design: config.py pins PROJECT = BRANCH = "_"; no auth, no multi-tenancy. Languages: TypeScript/TSX, Python, Java, Kotlin, Swift are SCIP-navigable; ~10 additional languages are search-only via Zoekt. Before 0.5.0 this project was called codeintel and shipped as codeintel-navigation-mcp (still installable at 0.4.0, unmaintained). Most modules are simplified ports of an internal multi-tenant reference implementation with authorization, project/branch dimensions, and async SQLAlchemy stripped out in favor of stdlib sqlite3 and the fixed (project='_', branch='_') pin — module docstrings name what was dropped and why.
Architecture & Data Flow
Indexing pipeline (index_repo() in index_cli.py, ~line 966): git gates (NotAGitRepositoryError) → scip version check (≥0.9.0; older scip silently writes a schema-valid db with zero chunks/mentions — verified regression) → language resolve (explicit --language else detect_language() counting extensions over git ls-files -z, ties by fixed priority) → language SCIP indexer subprocess (scip-typescript/-python/-java/-swift) to a scratch .scip file → scip expt-convert to a scratch index.db (documents/chunks/global_symbols/mentions/defn_enclosing_ranges schema) → graph.populate_graph_for_repo() reads the scratch db before publish to extract package deps into registry.db's packages/edges tables → git config zoekt.name <slug> pins the repo name, then zoekt-git-index runs against the working tree → optional semantic stage (chunker.chunk_file + embeddings.EmbeddingModel + semantic.SemanticStore, LanceDB) → index_cli._publish_atomically() write-temp-then-os.replace()s the current pointer file (the ONLY publish step) → registry.Registry.upsert() records the run outcome. On indexer failure, a degrade state machine (opt-in FALL-0x decisions) may publish search-only (Zoekt+semantic, no SCIP db, no current pointer) instead of leaving nothing — but only when not published yet; once the pointer has flipped, later failures are hard failures, never degrading an already-live good index.
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.
- 1mo ago Changed · +17 lines · +2,875 tokens per session 092b6118332d
- 1mo ago First seen · 106 lines · 2,927 tokens per session scan A 00f4b8d99797
jarvis AGENTS.md is an instructions file published in the GitHub repository jarvis-intelligence/jarvis (0 stars, last pushed 1mo ago), licensed MIT. It adds 5,802 tokens to every session, about $0.0232 per session on Opus 5.5. A static security scan graded it A with 1 finding (runs shell commands). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other instructions, from other repositories
repobrain copilot-instructions.md
Copilot instructions for study8677/repobrain, covering github copilot bootstrap instructions and hard rule — query the repobrain hub first.
vscode buildNext.instructions.md
Working notes and architecture documentation for the new esbuild-based build system in build/next. Use when making changes to the new build pipeline (transpile/bundle commands, NLS plugin, source-map handling, resource copying, or self-hosting watch tasks).
next.js AGENTS.md
AGENTS.md instructions for vercel/next.js, covering next.js development guide, codebase structure, monorepo overview, core package: packages/next and other important packages.
vscode oss-third-party-notices.instructions.md
Instructions for microsoft/vscode, covering vs code oss third-party-notices pipeline, architecture, pipeline flow in ci, cache eligibility and rollout and applying the notice (cutover).
langchain AGENTS.md
AGENTS.md instructions for langchain-ai/langchain, covering global development guidelines for the langchain monorepo, corridor security analysis, project architecture and context, monorepo structure and development tools & commands.
vscode copilot-instructions.md
Copilot instructions for microsoft/vscode, covering vs code copilot instructions, project overview, root folders, core architecture (src/ folder) and built-in extensions (extensions/ folder).