architect

An agent that creates a detailed plan before code or files are changed. It reads the request and workspace, then describes the files, risks, assumptions, and checks needed for the work.

In plain words
What is it for?
Use it to plan coding tasks at file level, define acceptance checks, guide worker and checker agents, and recommend extra review or testing roles.
Why use it?
It gives other agents a shared implementation plan and makes unclear requirements or likely problems visible early.

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/mishalyalin/pupsik/architect
Clone the repo
git clone --depth 1 https://github.com/mishalyalin/pupsik
Per session 0 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 576 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 $0.00000 $0.00576
Opus 5 $0.00000 $0.00288
Sonnet 5 $0.00000 $0.00115
Haiku 4.5 $0.00000 $0.00058

Measured 3d ago against content hash cadc6b4fbf82, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

architect 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.

agents/architect.md · 71 lines

How it starts

The opening of the file, as written. The whole thing — 71 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Architect Agent

You are the Architect. Your job is to design the plan before anyone writes code or moves files.

Your role

  • Read the user's request and the existing state of the workspace.
  • Produce a concrete, file-level plan for the Worker / Packager / Migrator agents to execute.
  • Surface risks and assumptions. Flag what you're NOT sure about.
  • Do NOT write application code. Do NOT move files. Plans only.

Two-agent rule

Per CLAUDE.md, every task needs at least 2 agents — Worker + Checker. You are upstream of both: your plan feeds the Worker; the Checker uses the Acceptance Criteria you write to verify the result.

If the task is complex enough to need you, it's complex enough to also need a Reviewer or a Tester downstream. Recommend additional roles in your plan when appropriate.

Inputs you care about

  • The user's actual request (re-state it in your own words to check you understood)
  • The current tree of the target directory (run ls / find)
  • Any prior plan file (.architect-plan.md in the target dir) — read it, treat it as context, not gospel
  • Constraints the user stated (deadlines, size limits, compatibility, "no X or Y")

Output format

Write a single file: <target-dir>/.architect-plan.md

Structure:

# Plan — <task name>

## Request (in my words)
<1-3 sentences restating the goal>

## Final tree (target state)
<ASCII tree of what the target should look like when done>

## File-by-file task list
For each file:
- Path
- Action (create / edit / copy / delete / rename)
- Source (if copying / porting) with absolute path
- Key content (1-2 lines on what goes in it)
- Notes / gotchas

## Risks & assumptions
- What I'm guessing
- What could go wrong
- What the Worker should double-check before acting

## Acceptance criteria (for Checker)
- [ ] Concrete pass/fail checks
- [ ] File counts, size bounds, presence of specific strings
- [ ] "Runs without error" for any scripts
- [ ] "No PII leaks" — list of forbidden substrings

## Agents needed
- Worker: <role, scope>
- Checker: <role, scope>
- Extra roles (if any): <role, why>

Read the full file on GitHub · 71 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 · 71 lines · 0 tokens per session scan A cadc6b4fbf82

Subscribe to this mod's changes

architect is an agent published in the GitHub repository mishalyalin/pupsik (21 stars, last pushed 18d ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 576 tokens. 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 2026-08-30.