Sruja

Sruja is a context engineering tool for software teams. It helps you:

• capture important knowledge and decisions
• retrieve the right task context for developers and AI agents
• verify that code changes still align with those decisions

The product center is not diagrams or generic agent orchestration. The product center is a small loop:

1. capture knowledge and reviewed intent
2. brief the next change with grounded context
3. verify the result with deterministic checks

Core Foundations

1. Capture

Bring durable context into the repo:

• sruja ingest <path> for design docs, ADRs, and notes
• sruja decision ... for decision records and links
• optional repo.sruja when you want reviewed intent in Git

2. Retrieve

Give humans and AI the minimum useful context before editing:

• sruja focus -r . --file path/to/file.rs
• sruja ai -r . --task "Refactor auth boundary"
• sruja mcp -r . for editor and agent tooling

3. Verify

Check that implementation still matches reality and intent:

• sruja drift -r . --structural-only --advisory
• sruja intent check -r .
• sruja verify-task --profile coding -r .
• sruja lint repo.sruja when reviewed intent exists

Quick Start

Install the CLI and run the core loop:

curl -fsSL https://sruja.ai/install.sh | bash

# 1) detect current structure
sruja start -r .
sruja drift -r . --structural-only --advisory

# 2) brief a change
sruja focus -r . --file path/to/file.rs

# 3) verify the result
sruja verify-task --profile coding -r .

This works without a repo.sruja file. Add reviewed intent later if you want stricter decision tracking in Git.

Optional Reviewed Intent

When you want durable, reviewable intent in version control:

sruja lint repo.sruja
sruja sync -r .
sruja drift -r . -a repo.sruja

Treat repo.sruja as reviewed truth, not as the day-one requirement.

Editor Integration

Sruja works best when your editor can ask for context directly through MCP.

Cursor template: .cursor/mcp.json

{
  "mcpServers": {
    "sruja": {
      "command": "sruja",
      "args": ["mcp", "-r", "."],
      "env": {
        "SRUJA_MCP_TOOL_PROFILE": "coding",
        "SRUJA_MCP_READONLY": "1"
      }
    }
  }
}

Use the default coding profile. Host editors own the LLM loop; Sruja supplies grounded context and verification.

Extensions

Everything else should build on the core foundations above.

Core-adjacent extensions

• architecture authoring: richer repo.sruja workflows, proposals, author evidence
• visualization: Mermaid, Markdown, D2, GraphML, Neo4j, Obsidian exports
• team workflows: review, drift in CI, compliance, critique, federation

Advanced extensions

• agent ops: plan/apply, learnings, run snapshots, memory curation
• workflow orchestration: workflow and AI-DLC flows
• inspection and analytics: dashboards, graph history, metrics, registry tooling

These remain useful, but they are not the product story to lead with.

Product Contract

Every public feature should strengthen one of these jobs:

• capture knowledge
• surface decisions
• retrieve relevant context
• verify alignment with intent

If a feature does none of those, it belongs in an extension or should be hidden.

Docs

• Website book: https://sruja.ai
• Getting started: book/src/docs/getting-started.md
• Host/editor setup: HOST_AGENT_INTEGRATION.md
• Core vs extensions: FEATURE_TIERS.md
• MCP tools reference: mcp_tools_reference.md
• Context framing: CONTEXT_ENGINEERING.md

Installation

CLI

Install script

curl -fsSL https://sruja.ai/install.sh | bash

Cargo

cargo install sruja-cli --git https://github.com/sruja-ai/sruja

Build from source

git clone https://github.com/sruja-ai/sruja.git
cd sruja
just build

Optional Skills

Harness skill:

npx skills add https://github.com/sruja-ai/sruja --skill sruja-harness

Architecture authoring skill:

npx skills add https://github.com/sruja-ai/sruja --skill sruja-architecture

Troubleshooting

sruja: command not found

export PATH="$HOME/.local/bin:$PATH"

AI generates invalid reviewed intent

sruja lint repo.sruja --format json

Need to understand what Sruja found

• Run sruja focus before a change
• Run sruja ai for a paste-ready task brief
• Run sruja why for investigation

Contributing

• Contributing Guide
• Roadmap

Contributors

This project follows the all-contributors specification. Contributions of any kind are welcome!

License

Apache 2.0 or MIT