research

A research workflow for understanding a coding task before planning it. It combines issue details, codebase analysis, dependency checks, repository exploration, and external documentation research.

In plain words
What is it for?
Use it with a GitHub issue number or a text topic. It can inspect the issue, trace affected code and dependencies, review architecture, and gather relevant documentation.
Why use it?
It reveals which files and symbols a change may affect and helps uncover existing design constraints. This lowers the chance of missing related code or relying on outdated assumptions.

Command for Claude Code

Install

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.

agentmods
npx agentmods add commands/cognitx-leyton/codegraph/research
Clone the repo
git clone --depth 1 https://github.com/cognitx-leyton/codegraph

Made for: Claude Code.

Per session 10 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 672 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.00010 $0.00672
Opus 5 $0.00005 $0.00336
Sonnet 5 $0.00002 $0.00134
Haiku 4.5 $0.00001 $0.00067

Measured 2d ago against content hash 9a62d6fa1076, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

research 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.

.claude/commands/research.md · 109 lines

How it starts

The opening of the file, as written. The whole thing — 109 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Research (Step 2)

Input: $ARGUMENTS

Deep research before creating an implementation plan. Combines codegraph queries, codebase exploration, and external documentation research.

Phase 1: Issue Context

If input is a number, fetch the GitHub issue:

gh issue view $ARGUMENTS

Extract: title, description, labels, linked PRs, comments.

If input is text, use it directly as the research topic.

Phase 2: Codebase Impact Analysis

Use codegraph to understand what the change touches:

2.1 Find affected symbols

/graph "MATCH (c:Class) WHERE c.name CONTAINS '<keyword>' RETURN c.name, c.file"
/graph "MATCH (f:Function) WHERE f.name CONTAINS '<keyword>' RETURN f.name, f.file"

2.2 Blast radius

For each affected symbol, check who depends on it:

/blast-radius <symbol>

2.3 Related files

/graph "MATCH (f:File)-[r:IMPORTS_SYMBOL]->(g:File) WHERE g.path CONTAINS '<path>' RETURN f.path, r.symbol"

2.4 Architecture check

/arch-check

Note any existing violations that might interact with the change.

Phase 3: Codebase Exploration

Use the Explore agent to:

  • Read the files identified in Phase 2
  • Understand existing patterns and conventions
  • Identify test files that cover the affected code
  • Note any related TODO/FIXME/HACK comments

Phase 4: External Documentation (if needed)

If the issue involves external libraries, APIs, or frameworks:

4.1 Context7 (for library docs)

Use context7 MCP to fetch current documentation for the relevant library.

4.2 NotebookLM (for deep research)

If complex research is needed:

nlm notebook list  # check existing notebooks
nlm notebook query <notebook-id> "<question>"

4.3 Web search (fallback)

Search for specific error messages, migration guides, or API references.

Phase 5: Research Summary

Output a structured summary:

## Research: <topic>

### Issue
<Issue title and key requirements>

### Affected Code
| File | Symbol | Impact |
|------|--------|--------|
| path | name | what changes |

### Blast Radius
- N files import the affected symbols
- N tests cover the affected code
- Key callers: <list>

### External Context
<Any library docs, API details, or migration notes>

### Risks
| Risk | Severity | Mitigation |
|------|----------|------------|
| ... | HIGH/MED/LOW | ... |

### Recommended Approach
<1-2 paragraphs on how to implement this>

### Ready for: /plan_local

Read the full file on GitHub · 109 lines

Changes

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.

  1. 2d ago First seen · 109 lines · 10 tokens per session scan A 9a62d6fa1076

Subscribe to this mod's changes

research is a command published in the GitHub repository cognitx-leyton/codegraph (11 stars, last pushed 1mo ago), licensed Apache-2.0. It adds 10 tokens to every session and 672 once invoked, about $0.0001 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 2026-08-30.