name: roam description: > Codebase comprehension via roam-code CLI. Use when exploring codebases, planning modifications, debugging failures, assessing PR risk, or checking architecture health. Triggers on: understanding project structure, pre-change safety checks, finding symbols/files, blast radius analysis, affected tests, health scoring, refactoring guidance, code review. Requires roam-code installed (pip install roam-code) and an indexed project (roam init).

Roam — Codebase Comprehension Skill

Repository: https://github.com/Cranot/roam-code

Roam pre-indexes codebases into a semantic graph (symbols, dependencies, call graphs, architecture layers, git history) stored in a local SQLite DB. Query it via CLI instead of repeatedly grepping files and guessing structure.

Setup

Ensure roam-code is installed and the project is indexed:

pip install roam-code   # or: pipx install roam-code
cd <project-root>
roam init               # indexes codebase, creates .roam/index.db

After git pull or major changes, run roam index to refresh (incremental, near-instant if few files changed). After large refactors: roam index --force.

Command Decision Table

Use this table to pick the right command for the situation:

Core Workflow

1. Orientation (first time in a repo)

roam understand        # tech stack, architecture, health, conventions
roam tour              # onboarding: key symbols, reading order, entry points
roam map               # project skeleton with top symbols by PageRank

2. Before Making Changes

Always run roam preflight <symbol> before modifying code. It combines blast radius + affected tests + complexity + coupling + fitness into one check:

roam preflight MyClass
# Output: blast radius, affected tests, complexity, coupling, fitness verdict

If you only need files to read:

roam context MyClass
# Output: definition file + callers + callees with exact line ranges

3. After Making Changes

roam diff              # blast radius of uncommitted changes
roam diff --staged     # blast radius of staged changes
roam pr-risk           # risk score (0-100) + suggested reviewers

4. Debugging

roam diagnose <symbol>  # root cause ranking by z-score risk
roam trace <A> <B>      # dependency path between two symbols
roam effects <symbol>   # DB writes, network I/O, filesystem, global mutation

Output Modes

• Default: Human-readable text (also optimized for LLM consumption)
• roam --json <cmd>: Structured JSON with consistent envelope
• roam --budget N <cmd>: Token-capped output (N = max tokens)
• roam --sarif <cmd>: SARIF 2.1.0 for CI integration

Prefer --json when you need to parse output programmatically. Prefer --budget 2000 when context window is tight.

Key Commands Reference

roam search <pattern>

Find symbols by name (regex). Results ranked by PageRank.

roam search "Auth.*Service"
roam search "handle_request" --kind fn

roam context <symbol>

AI-optimized file list with line ranges for reading. Supports --task modify|debug|review for context tuning.

roam context Flask
roam context myfile:MyFunction    # disambiguate with file prefix

roam preflight <symbol|file>

Compound pre-change check. Run this before every modification.

roam preflight UserController
roam preflight src/auth/login.py

roam health

Composite score (0-100). Use --gate score>=70 for CI.

roam health
roam health --gate score>=70    # exit 1 on failure

roam diff

Blast radius of uncommitted or committed changes.

roam diff                       # uncommitted
roam diff --staged              # staged only
roam diff HEAD~3..HEAD          # commit range

roam algo

Detect algorithm anti-patterns (23 patterns: O(n^2) loops, N+1 queries, quadratic string building, etc.) with confidence levels and fix suggestions.

roam algo
roam algo --confidence high     # high-confidence only
roam algo --task nested-lookup  # specific pattern

roam impact <symbol>

Full blast radius using Personalized PageRank.

roam impact Flask

roam symbol <name>

Symbol definition + callers + callees + metrics.

roam symbol open_db
roam symbol --full open_db      # include source code

roam affected-tests <symbol|file>

Trace reverse call graph to find covering tests.

roam affected-tests UserService

roam agent-export --write

Auto-generate agent instructions for the project. Detects CLAUDE.md, AGENTS.md, .cursor/rules, etc.

roam agent-export --write
roam agent-export --agent-prompt    # compact ~500-token prompt

roam minimap --update

Inject/refresh annotated codebase snapshot in CLAUDE.md.

roam minimap --update           # update sentinel block in CLAUDE.md

Discovering More Commands

This skill covers the most common commands, but roam has 144 commands. To explore what's available:

roam --help                 # list all available commands
roam <command> --help       # detailed usage for a specific command

For full documentation, examples, and the latest features, see the roam-code repository.

Tips

• One roam command replaces 5-10 grep/read cycles. Always try roam first.
• Use roam search instead of grep/glob for finding symbols — it understands definitions vs. usage and ranks by importance.
• roam context gives exact line ranges — more precise than reading whole files.
• After git pull, run roam index to keep the graph fresh.
• For disambiguation, use file:symbol syntax: roam symbol myfile:MyClass.