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.
npx agentmods add commands/alphaaiservice/cortex/auto-buildgit clone --depth 1 https://github.com/alphaaiservice/cortexWhat 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 | $0.00058 | $0.08249 |
| Opus 5 | $0.00029 | $0.04124 |
| Sonnet 5 | $0.00012 | $0.01650 |
| Haiku 4.5 | $0.00006 | $0.00825 |
Grade A, and why
auto-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 yesterday.
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.
How it starts
The opening of the file, as written. The whole thing — 609 lines — stays where its author put it; the contents beside it link to each section on GitHub.
🤖 Autonomous Product Builder (Auto-Build Loop)
You are an autonomous full-stack product builder following Alpha AI Service's engineering standards. You will take a product specification and build the ENTIRE product from scratch — without any human intervention — iterating until the product is complete, tested, and deployable.
Input: $ARGUMENTS
📥 INPUT PARSING — FIRST STEP (before anything else)
Parse $ARGUMENTS to determine the input file:
1. If $ARGUMENTS is EMPTY (user just typed /auto-build with nothing):
→ Auto-detect: Search the current directory for spec files in this priority order:
a. PRD.md
b. prd.md
c. SPEC.md / spec.md
d. DESIGN.md / design.md
e. PLANNING.md / planning.md
f. REQUIREMENTS.md / requirements.md
g. product-brief.md
h. Any single .md file in the directory (if only one exists besides README.md/CHANGELOG.md)
→ If found → use it as SPEC_FILE
→ If multiple found → list them and ask user to pick one
→ If NONE found → ask user: "No spec file found. Describe your product or provide a file path."
2. If $ARGUMENTS is a file path (ends with .md, .txt, .doc, .pdf, or exists on disk):
→ SPEC_FILE = that file path (e.g., ./design.md, ./planning-doc.md, ./PRD.md, ~/docs/spec.md)
→ Read the file content as the product specification
3. If $ARGUMENTS is a quoted string (not a file path):
→ Treat it as an inline product description
→ Generate a PRD from it first, save as PRD.md, then use that as SPEC_FILE
4. If $ARGUMENTS contains flags:
→ --lang python|nestjs|springboot → override language detection
→ --max-iterations N → set max iterations
→ --workflow → orchestrate the build with native Claude Code **Workflows** (background,
resumable, context-light) instead of the interactive Agent-Teams / sequential loop.
See PHASE 0.5 (Build Mode Selection) and `references/AUTO_BUILD_WORKFLOW.md`.
→ The remaining non-flag part is the file path or description
IMPORTANT: The input file can be ANY name — PRD.md, spec.md, design.md, planning.md,
requirements.md, product-brief.md, etc. Do NOT hardcode "PRD.md" as the expected filename.
Always use the actual filename from $ARGUMENTS or auto-detected file.
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.
- yesterday First seen · 609 lines · 58 tokens per session scan A 0d86c52edc02
auto-build is a command published in the GitHub repository alphaaiservice/cortex (1 stars, last pushed 25d ago), licensed MIT. It adds 58 tokens to every session and 8,249 once invoked, about $0.0003 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.
Other commands, from other repositories
alfred
Asistente contextual de Alfred Dev. Enruta automáticamente al flujo o comando operativo correcto.
feature
Ciclo completo de desarrollo: producto, arquitectura, desarrollo, QA, docs, entrega.
_composicion
Protocolo interno compartido para la composición dinámica del equipo de Alfred según tarea, stack y señales runtime.
ajustes
Configura Alfred Dev: autonomía, proyecto, Lucius, memoria y personalidad. Antes /alfred-dev:config.
audit
Auditoría completa del proyecto con 4 agentes en paralelo.
discuss
Refina una idea o feature antes de abrir un flujo completo de implementación.