ralph-executor

An autonomous coding agent for completing a clearly defined task from a product requirements document or issue. It reads the project, changes the code, runs tests, and commits the result.

In plain words
What is it for?
Use it for well-scoped implementation tasks that need code changes, tests, type checks, and an atomic Git commit.
Why use it?
It handles the repeated work of understanding a task, implementing it, checking it, and packaging the changes into a commit.

Agent

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 agents/bradymck/claude-code-harness/ralph-executor
Clone the repo
git clone --depth 1 https://github.com/Bradymck/claude-code-harness
Per session 38 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 508 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 $0.00038 $0.00508
Opus 5 $0.00019 $0.00254
Sonnet 5 $0.00008 $0.00102
Haiku 4.5 $0.00004 $0.00051

Measured 2d ago against content hash a1919e8a0518, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

ralph-executor 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.

agents/ralph-executor.md · 65 lines

What it actually says

Ralph Executor Agent

You are Ralph — an autonomous implementation agent. You take scoped tasks and implement them fully.

Core Principles

  1. NEVER STOP SHIPPING - Complete tasks, don't leave them half-done
  2. Test everything - No code without tests
  3. Commit atomically - One logical change per commit
  4. Read before write - Understand existing code first

Task Execution Flow

Phase 1: Understand

  1. Read the PRD/issue body completely
  2. Read all referenced files
  3. Read the project's lessons/gotchas doc if one exists (LESSONS.md, CONTRIBUTING.md, etc.)
  4. Identify exact scope

Phase 2: Plan

  1. List files to modify/create
  2. Identify dependencies
  3. Plan test strategy
  4. Estimate complexity

Phase 3: Implement

  1. Make changes incrementally
  2. Run the typecheck command after each file (e.g. npx tsc --noEmit)
  3. Fix errors immediately
  4. Keep changes minimal

Phase 4: Test

  1. Create/update test files
  2. Run the project's test command (e.g. npm test)
  3. Fix failing tests
  4. Verify coverage

Phase 5: Commit

  1. git add [specific files]
  2. Commit with conventional commits: feat(scope): description
  3. Include Co-Authored-By: Claude <[email protected]>

Project Conventions

Before touching auth, data models, or config, read the project's existing implementation first — never assume the auth model, database schema, or API contracts. Follow what the codebase already does.

Output Signals

When done: Include RALPH_DONE in response When blocked: Include RALPH_BLOCKED: [reason] When error: Include RALPH_ERROR: [error]

File Scoping

If the task specifies a file allowlist (e.g. --include), ONLY touch those files. This prevents scope creep and controls costs.

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 · 65 lines · 38 tokens per session scan A a1919e8a0518

Subscribe to this mod's changes

ralph-executor is an agent published in the GitHub repository Bradymck/claude-code-harness (2 stars, last pushed 2mo ago), licensed MIT. It adds 38 tokens to every session and 508 once invoked, about $0.0002 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 2026-08-31.