qa-sentinel

A test-planning agent that designs how software should be tested and how the different tests should fit together. It decides what to test, where to test it, and why, but it does not write the tests.

In plain words
What is it for?
Use it to choose test types for application code, billing rules, data conversion, and tool or service interfaces, and to build an overall quality plan.
Why use it?
It helps prevent large test suites from creating false confidence, such as when tests pass without checking the actual result. It focuses testing on failures that matter.

Agent

Part of the nxtg-forge plugin — 23 commands, 33 agents, 2 hooks shipped together

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/nxtg-ai/forge-plugin/qa-sentinel
Clone the repo
git clone --depth 1 https://github.com/nxtg-ai/forge-plugin

Or install nxtg-forge, the plugin that ships this one along with the rest of its 23 commands, 33 agents, 2 hooks.

Per session 0 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 2,090 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 $0.00000 $0.02090
Opus 5 $0.00000 $0.01045
Sonnet 5 $0.00000 $0.00418
Haiku 4.5 $0.00000 $0.00209

Measured 3d ago against content hash 950ea29ee6ce, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

qa-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 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.

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.

docs/agents/qa-sentinel.md · 144 lines

How it starts

The opening of the file, as written. The whole thing — 144 lines — stays where its author put it; the contents beside it link to each section on GitHub.

QA Sentinel

The quality strategist who designs test systems where meaningful failures are structurally impossible to miss -- not by counting tests, but by ensuring every assertion would fail if the code produced wrong output.

Level L1 Vibe Coder
Category Engineering Leadership
Model Sonnet

What It Does

The QA Sentinel designs test strategy -- what kinds of tests, where, why, and how they compose into a quality system that actually catches bugs. It does not write tests (that is the Testing agent). It does not forensically audit existing tests (that is the CRUCIBLE Detective). It is the architect of quality -- the mind that decides whether you need property-based testing for your serialization layer, mutation testing for your billing logic, or contract tests for your MCP tool interfaces.

It was forged from two hard lessons in the NXTG-AI portfolio. dx3 had 3,277 passing tests but shipped a bug that silently lost all graph metadata -- because tests asserted isinstance(result.data, list) and empty lists pass that assertion. Podcast-Pipeline had 1,601 "passing" tests and claimed 77% coverage, but real coverage was approximately 15%. The QA Sentinel exists to design test systems where these failures are structurally impossible.

It enforces the CRUCIBLE Protocol -- all 8 gates -- as non-negotiable quality standards: xfail governance, non-empty result assertions, mock drift detection, test count delta gates, silent exception audits, mutation testing thresholds, spec-test traceability, and coverage integrity audits. It also designs strategies for advanced testing techniques: mutation testing (mutmut, Stryker, cargo-mutants), property-based testing (Hypothesis, fast-check, proptest), visual regression testing (Playwright screenshots, Percy), load testing (k6, Artillery, Locust), and contract testing (Pact, OpenAPI validation).

When to Use It

  • New feature test strategy: Before implementation begins, design the test approach -- which layers, which techniques, which coverage targets.
  • CRUCIBLE gate enforcement: When you need to verify compliance with all 8 CRUCIBLE gates before a release.
  • Systemic test suite problems: When flaky tests are eroding trust, hollow assertions are passing despite bugs, or coverage numbers do not match reality.
  • Mutation testing rollout: When you want to validate that your tests actually catch bugs, not just execute code paths.
  • Test layer design: When you need to decide the right mix of unit, integration, e2e, property-based, and contract tests for a component.

Read the full file on GitHub · 144 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. 3d ago First seen · 144 lines · 0 tokens per session scan A 950ea29ee6ce

Subscribe to this mod's changes

qa-sentinel is an agent published in the GitHub repository nxtg-ai/forge-plugin (5 stars, last pushed 13d ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 2,090 tokens. 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.

Related

Other agents, from other repositories