An autonomous coding agent for the Ralph loop. In each iteration it takes one unfinished task, implements it, runs available checks, reviews the result, updates task status, and creates a focused Git commit.

Use it to implement tasks from Ralph task lists, run linting, type checks, or tests when available, record activity, verify acceptance criteria, and commit completed work.

It breaks extended development work into fresh, smaller work sessions and includes a separate review step instead of carrying all prior context forward.

ralph-agent — Agent for Claude Code by joaquimscosta

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.

Clone the repo

git clone --depth 1 https://github.com/joaquimscosta/arkhe-claude-plugins

Made for: Claude Code.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

README.md

[![agentmods](https://agentmods.dev/badge/agents/joaquimscosta/arkhe-claude-plugins/ralph-agent/github.svg)](https://agentmods.dev/agents/joaquimscosta/arkhe-claude-plugins/ralph-agent)

Your own site

<a href="https://agentmods.dev/agents/joaquimscosta/arkhe-claude-plugins/ralph-agent"><img src="https://agentmods.dev/badge/agents/joaquimscosta/arkhe-claude-plugins/ralph-agent/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

Your own site · 80×15

<a href="https://agentmods.dev/agents/joaquimscosta/arkhe-claude-plugins/ralph-agent"><img src="https://agentmods.dev/badge/agents/joaquimscosta/arkhe-claude-plugins/ralph-agent.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 52 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.

When invoked 1,210 The whole file, excluding the scripts and references it only reads on demand.

Security scan A 1 finding. A grade says what 26 rules found in the file — not that it is safe.

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.

Model	Per session	Once invoked
Fable 5.1	$0.00052	$0.01210
Opus 5.5	$0.00021	$0.00484
Sonnet 5.5	$0.00010	$0.00242
Haiku 4.5	$0.00005	$0.00121

Measured today against content hash a812150c70a9, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

ralph-agent scanned grade A with 1 finding 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 today.

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.

Makes network callslowCapability

Not a fault in itself. Listed so you know the mod talks to something, and to what.

| `api` | build checks + start server + `curl` endpoints + verify status/body + stop server |

plugins/ralph/agents/ralph-agent.md · 150 lines

How it starts

The opening of the file, as written. The whole thing — 150 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Ralph Autonomous Development Agent

You are an autonomous development agent operating within the Ralph Wiggum loop protocol. Each iteration you receive runs in a fresh context window to prevent context bloat and maintain focus.

Core Protocol

Hat-Lite System

You operate in two modes within each iteration:

BUILDER MODE (First ~65% of iteration)

Read .ralph/current-taskset/memories.md first for project patterns
Read .ralph/current-taskset/tasks.json to find the next incomplete task
Pick exactly ONE task where passes: false
Implement the task following the steps provided
Run verification commands (lint, typecheck, tests if available)
Document your work in .ralph/current-taskset/activity.log

VERIFIER MODE (Last ~35% of iteration)

Review the implementation you just completed
Verify it meets the task's acceptance criteria
Update the task status in .ralph/current-taskset/tasks.json
Create a single, focused git commit
(Optional) Save useful learnings to .ralph/current-taskset/memories.md
Decide: continue to next iteration or complete

Key Files

File	Purpose
`.ralph/current-taskset/tasks.json`	Source of truth for task state
`.ralph/current-taskset/activity.log`	Iteration history (append-only)
`.ralph/current-taskset/memories.md`	Persistent learnings across iterations
`.ralph/current-taskset`	Symlink to active task set
`PROMPT.md`	Your instructions (read at start)

Iteration Flow

1. ORIENT
   - Read .ralph/current-taskset/memories.md for patterns
   - Read last 5 entries from .ralph/current-taskset/activity.log
   - Read .ralph/current-taskset/tasks.json for current state

2. BUILD
   - Pick ONE task where passes=false
   - Implement the change (follow patterns from memories)
   - Run verification per the task's verificationTier

3. VERIFY
   - Review your implementation
   - Update task status in tasks.json
   - Append entry to activity.log
   - Commit changes with descriptive message

4. LEARN (Optional)
   - If you discovered something useful, append to memories.md
   - Categories: Patterns, Decisions, Fixes, Context

5. DECIDE
   - All tasks pass? → Output "RALPH_COMPLETE: All tasks verified"
   - Tasks remain? → End iteration (loop continues)

Read the full file on GitHub · 150 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.

today First seen · 150 lines · 52 tokens per session scan A a812150c70a9

Subscribe to this mod's changes

ralph-agent is an agent published in the GitHub repository joaquimscosta/arkhe-claude-plugins (21 stars, last pushed today), licensed MIT. It adds 52 tokens to every session and 1,210 once invoked, about $0.0002 per session on Opus 5.5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-10-16.

Building agents Commits Git and workflow ai-agents ai-developer-tools anthropic claude-code-agents claude-code-plugins

hydra-git

🟢 Hydra's git operations specialist. Handles all version control tasks: staging, committing with well-crafted Conventional Commits messages, branching, merging, rebasing, stashing, cherry-picking, log inspection, diff analysis, and conflict detection. Runs on the cheap tier — git operations are mechanical and…

AR6420/Hail_Hydra · 126 tokens

code-commit

Git commit specialist with deep knowledge of version control best practices and semantic commit messages. Automatically runs language-specific quality checks (formatters and type checkers) before committing code changes. Use when ready to commit changes to git.

RoniLeor/specWeaver · 47 tokens

gsd-executor

Executes GSD plans with atomic commits, deviation handling, checkpoint protocols, and state management. Spawned by execute-phase orchestrator or execute-plan command.

itsjwill/gsd-pro · 36 tokens

refactor-cleaner

Dead code cleanup and consolidation specialist. Use PROACTIVELY for removing unused code, duplicates, and refactoring. Runs analysis tools (knip, depcheck, ts-prune) to identify dead code and safely removes it.

affaan-m/ECC · 51 tokens

vibe-seed-pr-worker

Handles the git and GitHub side of a vibe-seed-refresh pull request in symphony-alpha for its orchestrator. Modes: open (push, PR from the template), push (push a fix the orchestrator committed and reply on its review threads), status (one-line PR state), sync (stage a merge of main for the orchestrator to commit)…

closedloop-ai/claude-plugins · 0 tokens

state-keeper

Use this agent to run sync reconciliation between fakoli-state's three sources of truth — the SQLite canonical state (.fakoli-state/state.db), the filesystem (packets/, .evidence-buffer/, worktrees), and git (branches, commits, claims). It surfaces drift as structured discrepancy reports — orphan branches (claimed…

fakoli/fakoli-plugins · 465 tokens

ralph-agent