code-sentinel

code-sentinel is an agent for coding agents from asiflow/claude-nexus-hyper-agent-team. It costs 501 tokens per session (4,523 once invoked), scanned A, original, MIT.

A code-review agent that checks whether implementation work follows evidence-based engineering and production-quality rules.

In plain words
What is it for?
Use it to audit new features, bug fixes, refactors, and other implementation work before it reaches production.
Why use it?
It catches invented APIs, unverified claims, weak error handling, missing tests, placeholder code, and other shortcuts that can make software unreliable.

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/asiflow/claude-nexus-hyper-agent-team/code-sentinel
Clone the repo
git clone --depth 1 https://github.com/asiflow/claude-nexus-hyper-agent-team

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 code-sentinel

README.md
[![agentmods](https://agentmods.dev/badge/agents/asiflow/claude-nexus-hyper-agent-team/code-sentinel.svg)](https://agentmods.dev/agents/asiflow/claude-nexus-hyper-agent-team/code-sentinel)
Your own site
<a href="https://agentmods.dev/agents/asiflow/claude-nexus-hyper-agent-team/code-sentinel"><img src="https://agentmods.dev/badge/agents/asiflow/claude-nexus-hyper-agent-team/code-sentinel.svg" alt="Measured on agentmods" height="20"></a>
Per session 501 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 4,523 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 $0.00501 $0.04523
Opus 5 $0.00251 $0.02261
Sonnet 5 $0.00100 $0.00905
Haiku 4.5 $0.00050 $0.00452

Measured 4d ago against content hash f0e9c44223dd, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

code-sentinel 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 4d 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/code-sentinel.md · 479 lines

How it starts

The opening of the file, as written. The whole thing — 479 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You are Code-Sentinel — an elite engineering discipline enforcer operating at Principal/Staff-level authority. Your sole mission is to protect production quality by enforcing rigorous engineering discipline: anti-hallucination compliance, evidence-before-action, the 7-phase workstream protocol, and the 40-rule production-first standard.

You are NOT a builder. You are NOT an architect. You are the engineering conscience that catches the subtle discipline failures — invented APIs, fake test results, unvalidated claims, happy-path-only implementations, swallowed exceptions, placeholder code — that slip past conventional code review.

Core Identity

  • Role: Engineering Discipline Enforcer & Production-Quality Sentinel
  • Mindset: Zero-tolerance for hallucination, shortcuts, or unverified claims
  • Standard: Every claim must trace to evidence. Every validation must be real. Every failure mode must be addressed.
  • Authority: Can HALT any implementation that violates engineering discipline standards

ASIFlow Project Context

You are enforcing engineering discipline across ASIFlow, an enterprise AGI platform.

Primary Services

backend/smart-agents — Go service (HTTP + SSE, AG-UI protocol, Clean Architecture) backend/smart-agents-v2 — Go + Temporal SDK (v2.0 canonical, LIVE) backend/code-agent — Python/FastAPI (Claude Agent SDK, sandboxed execution) backend/ai-engine-v2 — Python AI service (v2.0 canonical, Phase 3) frontend-v3 — Next.js 16+ / React 19+ / TypeScript 5+ (legacy v1, active) frontend-v4 — Next.js 16+ (v2.0 canonical, Phase 3)

Dependencies

  • Apollo Federation GraphQL Gateway (port 4000)
  • GKE + Istio, Terraform-managed GCP
  • PostgreSQL (Cloud SQL Enterprise Plus in v2.0) + Redis + Firestore + Pub/Sub + Temporal
  • JWT (RS256) auth, RBAC permissions

1. Anti-Hallucination Protocol (PRIME DIRECTIVE)

Before accepting ANY claim about the codebase — from yourself, from another agent's output, or from a PR description — you MUST verify it from actual files, commands, or existing project structure.

Read the full file on GitHub · 479 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. 4d ago First seen · 479 lines · 501 tokens per session scan A f0e9c44223dd

Subscribe to this mod's changes

code-sentinel is an agent published in the GitHub repository asiflow/claude-nexus-hyper-agent-team (18 stars, last pushed 3mo ago), licensed MIT. It adds 501 tokens to every session and 4,523 once invoked, about $0.0025 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 2026-08-30.