Skip to main content
subagent CX Codex

pseudocode subagent for Claude Code: what it does, how to install · ruvnet

Turns a finished spec into pseudocode: algorithms, data structures and a complexity estimate before any code is written.

no license ★ 850 updated 2026-10-26

Install into Claude Code · Manually: put the files in place All methods and platforms ↓
.claude/agents/pseudocode.md
Open on GitHub

What pseudocode does

The subagent covers one SPARC phase, Pseudocode. You hand it a specification; it returns algorithm blocks with declared inputs, outputs and a begin/end body, plus the data structures it settled on and what each operation costs. Its metadata marks it as an architect role with high priority.

The file also sets house style. How a login check should read in pseudocode, how to describe an LRU cache or a trie for permission checks, how a token bucket limiter looks step by step. Design patterns get a section of their own, and the output ends with a roadmap for implementation.

Two hooks wrap a run. On start it records the phase in memory and looks for a spec_complete entry; at the end it stores a completion marker. Both hooks call memory_store and memory_search, so if those commands are missing from your setup the hooks have nothing to run.

What it helps with

  • Turn a spec into algorithms before coding starts
  • Pick a data structure and see the cost of each operation
  • Estimate time and space complexity for a search path
  • Sketch a token bucket rate limiter in pseudocode
  • Write down a design pattern instead of holding it in your head

Example prompts for the agent

  • Take spec.md and write the Pseudocode phase: algorithms, data structures, complexity.
  • Draft pseudocode for user login with password verification and session creation.
  • Work out the complexity of catalog search and tell me what to change in the index.

How to install pseudocode

Manually: put the files in place

.claude/agents/pseudocode.md

For all projects use ~/.claude/agents/. Claude delegates by the description field or an @-mention

Step by step, with commands for each platform: Claude Code subagents: how to create one, call it and keep context clean

Questions about pseudocode

How do I install pseudocode in Claude Code?

Place the subagent file at .claude/agents/pseudocode.md.

What does pseudocode do?

It handles the Pseudocode phase of SPARC: turning a spec into algorithms, choosing data structures and weighing complexity.

What do the hooks need to work?

They run memory_store and memory_search: one saves the current phase, the other looks up a spec_complete record.

Similar subagents

All subagents for Agent setup: memory, context, orchestration →

pseudocode by other authors comes first.

pseudocode ruvnet A subagent for the Pseudocode stage of SPARC: it turns a finished spec into an algorithm draft and works out the complexity. CX Codex ★ 850repo stars sparc-implement ruvnet sparc-implement is a Claude Code skill that runs SPARC design phases, from algorithm pseudocode and module/API architecture to working code for SPARC development. CX Codex +1 735installs computer-science-algorithms pproenca computer-science-algorithms selects data structures and classical algorithms while flagging O(n²) bottlenecks in loops. Helps organize caching and streaming sketches. CX Codex +1 205installs typst-paper bahayonghang typst-paper — a skill for working on existing Typst manuscripts. Diagnoses builds, checks bibliographies, cleans grammar and AI traces in English and Chinese papers. CX Codex +1 1,600installs update-code-from-shorthand official github update-code-from-shorthand converts shorthand notes and pseudocode between edit markers into fully written code. It turns rough logic sketches into working implementations. CX Codex +1 ★ 41,430repo stars sparc-methodology ruvnet sparc-methodology — a skill for Claude Code, Codex, Cursor, Copilot, Gemini CLI and OpenCode: SPARC development methodology with 17 modes — from specification and architecture to TDD, code review and deployment. CX Codex +1 1,200installs vanity-engineering-review bencium vanity-engineering-review flags code and architecture built to impress rather than to ship. Useful before a refactor, a PR review or a rebuild debate. CX Codex +1 3,400installs

From the same set76

The whole set →
commands The commands command for Claude Code is a reference for AI pair programming sessions, managing driver and navigator roles and running tests and code review from the terminal. CX Codex +1 ★ 850repo stars dual-orchestrator dual-orchestrator routes work between interactive Claude Code and headless Codex: use one for reasoning and review, the other for parallel execution. Built for workflows that combine both. CX Codex ★ 850repo stars issue-tracker issue-tracker coordinates GitHub issue tracking with multi-agent swarms to structure bug reports, run progress updates, and sync project tasks. CX Codex +1 ★ 850repo stars tester tester — a subagent that designs and runs tests across the pyramid: units with mocks, integration against a real database, E2E flows, performance checks, and security validation. Keeps coverage above 80% and catches bugs early. CX Codex ★ 850repo stars goal-planner goal-planner — a GOAP planner subagent that turns a goal into an ordered set of actions and reworks the plan when a step fails. CX Codex ★ 850repo stars v3-memory-specialist The v3-memory-specialist subagent for Claude Code migrates scattered project memory into AgentDB with HNSW indexing, speeding semantic search and sharing context across agents. CX Codex ★ 850repo stars config Loads a reference for pair programming settings: driver, navigator and switch modes, verification thresholds, test runs and agent profiles kept in .claude-flow/pair-config.json. CX Codex +1 ★ 850repo stars workflow-automation workflow-automation integrates multi-agent swarms with GitHub Actions pipelines to generate CI configurations, select tests, and inspect build failures. CX Codex +1 ★ 850repo stars

Project sponsors

Companies that keep this analytics open