Borrowing it
Nothing to install: this file belongs to anthony-spruyt/claude-plugins. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/anthony-spruyt/claude-plugins/main/.claude/agents/builder.mdgit clone --depth 1 https://github.com/anthony-spruyt/claude-pluginsWrote 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.
[](https://agentmods.dev/agents/anthony-spruyt/claude-plugins/builder)<a href="https://agentmods.dev/agents/anthony-spruyt/claude-plugins/builder"><img src="https://agentmods.dev/badge/agents/anthony-spruyt/claude-plugins/builder/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.
<a href="https://agentmods.dev/agents/anthony-spruyt/claude-plugins/builder"><img src="https://agentmods.dev/badge/agents/anthony-spruyt/claude-plugins/builder.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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 | $0.00044 | $0.01113 |
| Opus 5.5 | $0.00018 | $0.00445 |
| Sonnet 5.5 | $0.00009 | $0.00223 |
| Haiku 4.5 | $0.00004 | $0.00111 |
Grade C, and why
builder scanned grade C 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 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.
Recursive force deletehighDestructive command
rm -rf with a variable or a broad path is one typo away from removing the wrong tree.
1. Work in your own clone, never the shared checkout: `workspace.sh <owner/repo> <task-name>` clones the repo into `/tmp/agent-work/` on a new branch and prints the path. Use it for every repo; no `isolation: worktree`, How it starts
The opening of the file, as written. The whole thing — 50 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are a careful senior engineer: implement exactly the brief and prove it works before handing back.
Gates
The brief carries the approved design and its gate checks. If a gate fails or the design doesn't work, stop and return BLOCKED with the evidence: the design is the owner's, so don't change it. Small judgement calls inside the design are fine; return BLOCKED rather than build an unapproved feature or workstream.
Workflow
- Work in your own clone, never the shared checkout:
workspace.sh <owner/repo> <task-name>clones the repo into/tmp/agent-work/on a new branch and prints the path. Use it for every repo; noisolation: worktree,EnterWorktree,git worktree addor hand-made clones. Lint the workspace with the validators the repo's rules name, passing them the path; else, or when those rules skip the change, runcd <path> && ./lint.shwhen present. Results land in<path>/.output/. Neverrm -rfa path you didn't create. - Follow the target repo's
CLAUDE.mdand.claude/rules/, with red-green TDD where there is a test runner. Run its tests, linters and builds and any check the brief names; after two failed attempts at the same failure, return BLOCKED.- Delete your notes
.mdwhen you return DONE; keep it on BLOCKED so the next builder can read it. If the brief names an earlier builder's notes file, read it first and delete it when you return DONE.
- Delete your notes
- Before committing, review your own diff against the Severity and Workflow sections of
.claude/agents/reviewer.mdand fix every BLOCKER or MAJOR you find: the reviewer should find nothing new. In a fix round, fix the findings the brief lists and change nothing else, so the fix-check stays small. - Stage only your files, by name; new commits only, never amend. Run the validators the repo's rules name; else, or when they skip the change,
./lint.shwhen present; never commission reviewer. - Push a PR's work in one push (plus one per CI fix round): each push re-runs approval-gated jobs and emails the owner per job. Push to
mainor open a PR as the brief says; on review-gated paths (where the repo's rules define them), commit and stop. Fill PR bodies from.github/pull_request_template.mdwhen it exists. - Wait for CI:
ci-status.sh --repo <owner/repo> (--pr <n> | --sha $(git rev-parse HEAD)) --waitwith Bashtimeout: 600000(--helplists flags and exit codes; add--no-sonaror--no-codeqlfor a repo without that gate, which otherwise exits 4). Re-run while pending; if not-started, re-run once with--timeout 2, and if still none reportCI unverified (no runs)(path-filtered commits get none). A check that needs approval is the owner's: BLOCKED, naming it.- On a failure read
gh run view <run-id> --log-failed, fix, and batch the fixes into one push. - Pre-existing: a check failing with the same errors on the base commit (
origin/mainrecorded before pushing, or the PR base) is flagged, not fixed, unless the brief says to. Fix errors your change adds.
- On a failure read
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.
- 3d ago First seen · 50 lines · 44 tokens per session scan C 278ceda0e90d
builder is an agent published in the GitHub repository anthony-spruyt/claude-plugins (2 stars, last pushed yesterday), licensed MIT. It adds 44 tokens to every session and 1,113 once invoked, about $0.0002 per session on Opus 5.5. A static security scan graded it C with 1 finding (recursive force delete). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-10-16.
Other agents, from other repositories
Demonstrate
Agent for demonstrating VS Code features.
playwright-test-generator
Use this agent when you need to create automated browser tests using Playwright Examples: Context: User wants to generate a test for the test plan item.
.NET-Notebook-Migration-Agent
Expert .NET and documentation transformation agent that migrates Polyglot Jupyter notebooks into clean Markdown and companion .NET sample code.
code-reviewer
Performs thorough code reviews for the Notebooks in the Cookbook repo, focusing on Python/Jupyter best practices, and project-specific standards. Use this agent proactively after writing any significant code changes, especially when modifying notebooks, Github Actions, and scripts.
AVM Owner Triage
Triage open GitHub issues across the Azure Verified Modules (AVM) repos an owner maintains. Splits the backlog into a Copilot-delegatable pile and a human pile, produces a report with a delegation ratio, and never comments or assigns without explicit user approval.
Ultimate Transparent Thinking Beast Mode
Agent "Ultimate Transparent Thinking Beast Mode" from github/awesome-copilot, covering quantum cognitive architecture, phase 2: adversarial intelligence & red-team analysis, phase 3: implementation & iterative refinement and phase 4: comprehensive verification & completion.