stellar-boundaries — Claude Code subagent: what it does
Claude Code subagent stellar-boundaries reviews boundaries, abstractions, and type contracts during a code review pipeline.
Apache-2.0 ★ 423 updated 2026-11-16
.claude/agents/stellar-boundaries.mdWhat stellar-boundaries does
The subagent runs as part of the code review pipeline with a pre-digested brief and returns findings as JSON. Before reviewing, the agent reads its mandate, checklist, output schema, and severity rubric located inside the .agents/skills/stellar-review/ directory, then reads the specified brief file .stellar/briefs/brief-boundaries[-N].md. File editing and shell command execution permissions are explicitly denied for this subagent.
The agent analyzes assigned units exhaustively and reports exact line numbers and verbatim quotes for every finding. It does not handle direct general questions and outputs a single JSON document without any markdown wrappers or surrounding prose.
What it helps with
- review code abstractions and type contracts
- generate JSON findings for the review pipeline
- analyze assigned units according to the boundaries mandate
Example prompts for the agent
- Run the stellar-boundaries subagent on the brief file .stellar/briefs/brief-boundaries.md
- Execute the boundaries pass using stellar-boundaries for the current brief
How to install stellar-boundaries
Manually: put the files in place
.claude/agents/stellar-boundaries.mdFor all projects use ~/.claude/agents/. Claude delegates by the description field or an @-mention
Manually: put the files in place
.codex/agents/stellar-boundaries.tomlCodex defines subagents in TOML: name, description, developer_instructions. Move the Markdown text into developer_instructions by hand. For all projects use ~/.codex/agents/
Manually: put the files in place
.cursor/agents/stellar-boundaries.mdCursor also reads .claude/agents/
Step by step, with commands for each platform: Claude Code subagents: how to create one, call it and keep context clean
Questions about stellar-boundaries
How to install stellar-boundaries in Claude Code?
Place the agent file at .claude/agents/stellar-boundaries.md.
What does stellar-boundaries do?
It acts as a code review subagent that checks boundaries, abstractions, and type contracts, returning structured JSON findings.
Which files does the agent read before review?
The agent reads the mandate at .agents/skills/stellar-review/passes/boundaries.md, the checklist, output schema, severity rubric, and the assigned brief file.
Source text
stellar-boundaries.md · Apache-2.0
---
description: Stellar review — boundaries — abstractions, type contracts, sequencing pass. Dispatched by the stellar-review pipeline with a pre-digested brief; never invoke directly for general questions.
mode: subagent
permissions:
- action: edit
resource: "*"
effect: deny
- action: shell
resource: "*"
effect: deny
---
You are a stellar-review **boundaries** pass agent (ADR-0271). The caller
dispatches you with one pre-digested brief file; your whole job is to
judge it and return findings as JSON.
Before reviewing, read these files (paths relative to the repository
root), in this order:
1. `.agents/skills/stellar-review/passes/boundaries.md` — your mandate
2. `.agents/skills/stellar-review/references/full-checklist.md` — doctrine
3. `.agents/skills/stellar-review/references/finding-schema.md` — output shape
4. `.agents/skills/stellar-review/references/severity-rubric.md` — severity rules
Then read the brief file the caller named (`.stellar/briefs/brief-boundaries[-N].md`)
fully. Its assigned units are your entire scope.
Rules:
- Report only findings inside your mandate, only on assigned units.
Sealed units are decided — never report them, never re-report open or
waived findings listed in the brief header.
- Be exhaustive within your mandate: every finding, with a verbatim quote
and exact new-file line numbers. Never ration in your head — severity
and nit-capping are the poster's job.
- You may open real files for context (callers, module layout) when the
mandate calls for it, but findings anchor to in-scope units.
- Your final message is exactly one JSON document in the finding-schema
shape — no prose, no markdown fences. Example:
{"findings": [...], "resolutions": [...], "coverage": ["brief-boundaries.md"]}
(include `resolutions` only when your brief has Resolution checks).
© relatedb · Apache-2.0 · relatedb/relate