build

A command that guides a coding task from deciding what to change through implementation and self-review, then returns the work for your approval.

In plain words
What is it for?
Use it to plan and build a feature or fix, review the riskiest parts, and stop before the change is shipped.
Why use it?
It keeps planning, test-first development, and review in one predictable process while leaving the final shipping decision with you.

Command

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 commands/prabhdeepsingh/claude-plugins/build
Clone the repo
git clone --depth 1 https://github.com/PrabhdeepSingh/claude-plugins
Per session 105 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 3,294 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 $0.00105 $0.03294
Opus 5 $0.00053 $0.01647
Sonnet 5 $0.00021 $0.00659
Haiku 4.5 $0.00011 $0.00329

Measured 2d ago against content hash 213638eda0e6, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

build 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.

sonu/commands/build.md · 100 lines

How it starts

The opening of the file, as written. The whole thing — 100 lines — stays where its author put it; the contents beside it link to each section on GitHub.

/build — decide → build → hand back

Contract: this command sequences the implementation lifecycle — design, build, self-review — and then hands back. It pauses once (for you to approve the design — or zero times when /sonu:factory routes a ticket whose spec you already approved) and stops once (at a green suite, before shipping). It never commits, never merges. It delegates to existing skills; it does not re-implement them — and it is additive, not a replacement: /sonu:design-tree, /sonu:tdd, and /sonu:ship work identically when run standalone. Run /sonu:ship when you're ready to ship.

Apply this to $ARGUMENTS — the text typed after the command. If that token appears literally or is empty, derive the task from the current discussion in context.

Delegation disposition (optional flag). $ARGUMENTS may carry one flag that sets model-tiering's disposition for this run up front, so the fan-out decision is yours rather than a mid-build judgment call. Strip the flag token from the task text before deriving the task:

  • --orchestrate — pre-authorize fan-out. At Phase 1 step 4, tag every step that clears model-tiering's four criteria — where the author would otherwise leave such a step in-session out of habit or a vague hunch, tag it instead. It does not relax the rule that doubt about any of the four criteria removes the tag: a step that doesn't clearly clear all four still stays in-session. (On the trivial path the flag is moot — Phase 1 is skipped, so no plan exists to tag, and a ≤~10-line change is cheaper to type than to route; note that and build inline.)
  • --solo — keep everything in-session. Skip model-tiering tagging entirely; Phase 2 builds inline.
  • neither — model-tiering's own balanced judgment (the default).

The disposition is a tie-breaker, not an override: --orchestrate never delegates model-tiering's Section 4 categories (design, integration, debugging, security, review-of-delegated-output) and cannot manufacture an executor tier — on a session with no trustworthy tier below it, it still no-ops and builds inline. --solo always wins: nothing is delegated regardless of tier.

If both --orchestrate and --solo are given, --solo wins — keeping work in-session is the safe direction. A repeated single flag (e.g. --orchestrate --orchestrate) is just that flag, deduped — not a conflict. Ignore any other flag-like token in the task text; only these two are recognized.


Read the full file on GitHub · 100 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. 2d ago First seen · 100 lines · 105 tokens per session scan A 213638eda0e6

Subscribe to this mod's changes

build is a command published in the GitHub repository PrabhdeepSingh/claude-plugins (3 stars, last pushed 2d ago), licensed MIT. It adds 105 tokens to every session and 3,294 once invoked, about $0.0005 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.