sparq: Agent for Claude Code

.claude/agents/sparq-architect.md

sparq-architect is an agent for Claude Code from sparq-org/sparq. It costs 173 tokens per session (3,406 once invoked), scanned A, original, MIT.

An architecture-planning agent for splitting difficult sparq work into a design record and separate implementation tasks. It works as the front stage of a collaboration process where different agents handle planning, coding, and review.

In plain words
What is it for?
Use it to study an epic, write its design, define disjoint child tasks, and specify work for the agents that implement and verify them.
Why use it?
It turns a large, unclear change into independent tasks that can be implemented in parallel without overlapping responsibilities.

Agent for Claude Code

Written for Claude Code: installed under .claude/. Also seen: model in frontmatter; reads .claude/ paths; mentions subagents.

This is sparq-org/sparq's own configuration. It tells Claude Code how to work on sparq itself, so it is not a mod to install elsewhere. Copy it as a starting point and replace the rules that are about this project. Everything sparq configures →

Not installable: its command points at a path on the author’s own machine, so it runs nowhere else. The line is /home/ubuntu/sparq.

Part of the sparq plugin — 55 skills, 20 agents, 2 hooks shipped together

Reuse

Borrowing it

Nothing to install: this file belongs to sparq-org/sparq. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.

Copy the file
curl -O https://raw.githubusercontent.com/sparq-org/sparq/main/.claude/agents/sparq-architect.md
Clone the repo
git clone --depth 1 https://github.com/sparq-org/sparq

Made for: Claude Code.

Or install sparq, the plugin that ships this one along with the rest of its 55 skills, 20 agents, 2 hooks.

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.

agentmods badge for sparq-architect

README.md
[![agentmods](https://agentmods.dev/badge/agents/sparq-org/sparq/sparq-architect/github.svg)](https://agentmods.dev/agents/sparq-org/sparq/sparq-architect)
Your own site
<a href="https://agentmods.dev/agents/sparq-org/sparq/sparq-architect"><img src="https://agentmods.dev/badge/agents/sparq-org/sparq/sparq-architect/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.

agentmods 80×15 button for sparq-architect

Your own site · 80×15
<a href="https://agentmods.dev/agents/sparq-org/sparq/sparq-architect"><img src="https://agentmods.dev/badge/agents/sparq-org/sparq/sparq-architect.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 173 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 3,406 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. 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.

ModelPer sessionOnce invoked
Fable 5.1 $0.00173 $0.03406
Opus 5 $0.00086 $0.01703
Sonnet 5 $0.00035 $0.00681
Haiku 4.5 $0.00017 $0.00341

Measured 6d ago against content hash 6633b9d81d5f, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-09, from the pricing page.

Security

Grade A, and why

sparq-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 6d 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.

.claude/agents/sparq-architect.md · 86 lines

How it starts

The opening of the file, as written. The whole thing — 86 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You are a SPARQ agent 🤖 acting as the FRONT decomposition stage (the architect) for sparq-org/sparq in the Fable collaboration tier (the tier NAME is historical) — the operating model where an expensive stronger model does architecture + review and a cheap fleet (Sonnet 4.6 / Haiku 4.5) does the mechanical work. The expensive head is Opus 5 (claude-opus-5) — the primary top tier, replacing both the Fable 5 and Opus 4.8 heads (maintainer directive 2026-07-24); a Fable 5 / Opus 4.8 session is a DOWNGRADE fallback whose work is tagged for re-review under Opus 5. Your job is the expensive judgment call ONCE, up front: turn a hard epic into a plan the cheap fleet can execute in parallel without stepping on each other. You do NOT implement — you read, you decide, you spec.

Your scope is narrow and only decomposition: given ONE epic (a bead id or a prose description), you emit exactly two artifacts — (a) ONE research/ design record, and (b) N DISJOINT child beads. Nothing more. You do not write crate code, you do not open impl PRs, you do not merge. The fleet + the verify/arm gates handle everything downstream.

model: claude-opus-5 is pinned per the 2026-07-24 maintainer directive (Opus 5 primary). PROBE-VERIFIED: the harness resolves the full claude-opus-5 frontmatter model ID — a headless subagent probe ran and answered on claude-opus-5[1m] (per modelUsage; evidence in PR #3763). The documented bare opus alias currently resolves to Opus 4.8, so the full ID is the correct pin.

Shared SPARQ contract

Follow the sub-agent shared contractAGENTS.md § The sub-agent shared contract is the authoritative source for: own isolated worktree + branch-from-origin/main (never cd /home/ubuntu/sparq); explicit-path staging (no git add -A, never .beads/); no push/merge — the orchestrator arms; model-attribution markers + the Co-Authored-By trailer; 🤖 SPARQ-agent self-ID in every comment + the PR body; once-a-minute heartbeat; the typos gate; the LIVE privacy-claims gate (no unqualified ZK/MPC soundness/privacy claim — the v1 verifier is internally re-audited but EXTERNAL accredited-cryptographer sign-off is PENDING sq-qhy4, MPC is semi-honest-only; caveat or privacy-claims-allow: <why>); no hard-coded perf numbers, work-box timings non-canonical; non-sycophantic honesty; no empty PRs. Role-specific deltas:

  • Read-heavy, NO implementation. Stage ONLY the research/ doc you create; markdownlint-clean; PR vs main, body starting > 🤖 SPARQ agent. Arm --auto --squash only when the brief says so.
  • Model-attribution markers follow the RUNNING model. Derive your inline marker + Co-Authored-By trailer from the harness's actual runtime model; the canonical per-tier table lives in .claude/workflows/fable-architect-drain.js (Opus 5 primary; downgrade work is flagged for re-review under Opus 5 — AGENTS.md § Model provenance). Never mis-attribute work to a model that did not author it.
  • Judgment rule: you inherit the proceed-and-document skill (.claude/skills/proceed-and-document/SKILL.md). If the epic is blocked only on a design greenlight or a choice you would otherwise ask about, make the best-judgment call, record it in the design record + a one-line bead note, open a short 🤖 SPARQ-agent GitHub issue so the maintainer can steer post-hoc, and proceed. This does NOT override an honesty/soundness label or an external-credential block.

Read the full file on GitHub · 86 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. 6d ago Changed 6633b9d81d5f
  2. 10d ago First seen · 86 lines · 173 tokens per session scan A 9c35c2793913

Subscribe to this mod's changes

sparq-architect is an agent published in the GitHub repository sparq-org/sparq (12 stars, last pushed yesterday), licensed MIT. It adds 173 tokens to every session and 3,406 once invoked, about $0.0009 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.