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/mutharasuarchunan13/codebase-intel/claude-mdgit clone --depth 1 https://github.com/MutharasuArchunan13/codebase-intelWhat 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 | $0.01248 | $0.01248 |
| Opus 5 | $0.00624 | $0.00624 |
| Sonnet 5 | $0.00250 | $0.00250 |
| Haiku 4.5 | $0.00125 | $0.00125 |
Grade A, and why
codebase-intel CLAUDE.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.
How it starts
The opening of the file, as written. The whole thing — 104 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Codebase Intel — Project CLAUDE.md
Project Overview
Codebase Intelligence Platform: an open-source, agent-agnostic system that provides AI coding agents with structured context, decision provenance, and quality contracts. Solves the three biggest gaps in AI-assisted development: context/memory, judgment/business context, and quality assurance.
Tech Stack
- Language: Python 3.11+
- Type System: Pydantic v2 for all models, mypy strict mode
- Storage: SQLite via aiosqlite (zero-dependency, portable)
- Parsing: tree-sitter for language-agnostic AST analysis
- Agent Interface: MCP (Model Context Protocol) server
- CLI: Typer + Rich
- Git Integration: GitPython
- Token Counting: tiktoken
- Hashing: xxhash for content fingerprinting
Architecture
Layered system with five core modules + two interface layers:
AI Agent (any) → MCP Server / CLI
↓
Context Orchestrator
↙ ↓ ↘
Code Graph Decisions Contracts
↘ ↓ ↙
Drift Detector
↓
Codebase (git)
Module Responsibilities:
- core/ — Shared types, config, exceptions. No business logic.
- graph/ — Semantic code graph: AST parsing, dependency mapping, impact analysis, SQLite storage
- decisions/ — Decision journal: structured records, git mining, code linking, temporal validation
- contracts/ — Quality contracts: architectural rules, pattern libraries, evaluation engine
- orchestrator/ — Context assembly: budget management, freshness scoring, conflict detection
- drift/ — Drift detection: staleness, pattern violations, knowledge decay
- mcp/ — MCP server: exposes all modules as queryable tools for AI agents
- cli/ — CLI interface: init, analyze, query, serve commands
Project Flow
codebase-intel init→ scans repo, builds initial code graph, generates starter configs- Git hooks keep graph updated incrementally on each commit
- AI agent connects via MCP → sends task description
- Orchestrator assembles relevant context (files, decisions, contracts) within token budget
- Agent receives structured context, writes code
- Drift detector flags violations post-commit
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 · 104 lines · 1,248 tokens per session scan A 26e3ecac54dd
codebase-intel CLAUDE.md is an instructions file published in the GitHub repository MutharasuArchunan13/codebase-intel (5 stars, last pushed 11d ago), licensed MIT. It adds 1,248 tokens to every session, about $0.0062 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
adrkit AGENTS.md
Instructions for mbeacom/adrkit, covering adrkit, visual adr graph, inbound @adr markers (v0.5.0: explain, check, and ci), the spec kit extension (packages/adapters/spec-kit) and adr queue.
adrkit copilot-instructions.md
Instructions for mbeacom/adrkit, covering adrkit — github copilot and github copilot specifics.
adrkit CLAUDE.md
Instructions for mbeacom/adrkit, covering adrkit — claude code and claude code specifics.
adr copilot-instructions.md
Instructions for zircote-plugins/adr, covering copilot instructions, project overview, key components, plugin structure and development guidelines.
adr-kit CLAUDE.md
Instructions for rvdbreemen/adr-kit, covering adr kit, backlog.md, werken in deze repo and graphify.
docflow AGENTS.md
Instructions for EvolveHQ/docflow, covering agents.md, what this repository is, repository structure, hard rules when editing adrs and domain-specific hard rules.