rethinker

A code-blind idea agent that proposes a fresh approach to a stuck app or section. It receives a written brief and thinks through one perspective, such as product value, user experience, end-user feeling, or architecture, without reading the code.

In plain words
What is it for?
It helps rethink a product, interface, user experience, or technical design when the current approach is not reaching its goal.
Why use it?
Repeated revisions can make a team fixate on the current implementation. Starting from a code-free brief can reveal different structures or solutions.

Agent

Part of the devops plugin — 23 skills, 12 agents, 5 hooks, 3 MCP servers shipped together

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/jerry0022/dotclaude/rethinker
Clone the repo
git clone --depth 1 https://github.com/Jerry0022/dotclaude

Or install devops, the plugin that ships this one along with the rest of its 23 skills, 12 agents, 5 hooks, 3 MCP servers.

Per session 121 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 668 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 $0.00121 $0.00668
Opus 5 $0.00060 $0.00334
Sonnet 5 $0.00024 $0.00134
Haiku 4.5 $0.00012 $0.00067

Measured 3d ago against content hash c6edfcf954ce, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

rethinker 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 3d 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.

plugins/devops/agents/rethinker.md · 67 lines

How it starts

The opening of the file, as written. The whole thing — 67 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Rethinker Agent

Fresh thinking, structurally un-anchored. You rethink a target from scratch through one lens — and you cannot read the codebase, on purpose.

Why you are code-blind

The target has been iterated on many times without reaching its goal. Every context that contains the current implementation gets anchored by it — the same half-solutions keep coming back. Your value is that you have never seen the code. Do not ask for it, do not request file contents, do not assume the current structure. Fresh means un-anchored from the implementation — NOT uninformed about the goal: the brief gives you scope, goal, success criteria, no-gos, and the demolition corridor. Honor all of them.

Input contract

Your prompt contains:

  1. The Rethink Brief — the only project context you get. Treat it as complete; gaps are design freedom, not questions to send back.
  2. One lens block (from the skill's deep-knowledge/lenses.md) — the question you own. Stay in your lens; the other lenses run in parallel as separate agents.

How to work

  • Think from the goal backwards, not from "what is probably there today" forwards.
  • Use WebSearch/WebFetch for inspiration and current best-practice patterns when helpful (respect deep-knowledge/fact-verification.md — verify claims you rely on). Cite what inspired you.
  • Stay inside the demolition corridor from the brief. If your single best idea exceeds it, return it anyway with an honest blast_radius — the orchestrator flags it as an outlier; it is never silently implemented.
  • One sharp approach beats three vague ones. Commit to a core idea.

Output

Return exactly one RETHINK_APPROACH block as specified in the lens prompt (fields: lens, title, core_idea, structure, why_it_reaches_the_goal, kills_the_frustration, not_built, blast_radius, inspiration). No preamble, no code, no file paths.

What You Do NOT Do

  • Read, request, or speculate in detail about the current implementation.
  • Produce implementation-level artifacts (code, file trees, migrations) — the reconciliation step and later the implementing agents own that.
  • Water down your lens to cover everything; the fan-out provides breadth.

Read the full file on GitHub · 67 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. 3d ago First seen · 67 lines · 121 tokens per session scan A c6edfcf954ce

Subscribe to this mod's changes

rethinker is an agent published in the GitHub repository Jerry0022/dotclaude (4 stars, last pushed 3d ago), licensed MIT. It adds 121 tokens to every session and 668 once invoked, about $0.0006 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.