Overview

At CIS, we do not treat an indicator as “done” just because it compiles or looks good on one chart. Every production script is held against a structured 30-point audit that checks the code, the math, the source behavior, the chart-context behavior, the debug output, the documentation, and the final release state before it is treated as ready.

What That Means in Practice

Every audit is run in one of two clearly labeled modes: Static source audit or Live chart validation. Every item in the checklist is scored PASS, WARN, or FAIL, and each score has to be backed by evidence, tested contexts, and notes. At the end, the audit must also produce a Status summary, Release verdict, and Primary fix target.

That structure matters because it keeps the process honest. A PASS needs evidence. A WARN is only allowed when there is an intentional, documented exception or an explicitly recorded live-validation deferral. A FAIL blocks release until it is fixed. In other words, the audit is not decorative. It is a release gate.

What the 30 Points Actually Cover

The checklist is broad by design. It starts with the fundamentals: does the script compile cleanly, follow suite structure, preserve settings stability, and stay aligned with the shared engine rules that govern routing, normalization, contract handling, data defense, market-type selection, and fallback behavior.

From there, it moves into the parts that are easiest to get wrong and hardest for a casual user to spot: math safety, warmup behavior, first-valid-bar honesty, reset behavior, plotted-source correctness, and the difference between a debug table that looks right and a production plot that is actually right. That distinction is explicitly part of the standards.

It also checks the finishing work that often gets skipped elsewhere: debug-table integrity, performance discipline, restoration after diagnostics, visual consistency, documentation accuracy, and final release discipline. The goal is not just to make the script work. The goal is to make sure the final version is correct, stable, explainable, and production-clean.

Static Source Audit vs. Live Chart Validation

A Static source audit is a source-level review. It checks code paths, settings, safeguards, standards alignment, and documented behavior, but it does not pretend that a full live chart matrix was executed if that did not happen. A Live chart validation goes further and records actual chart contexts, settings, reset modes, and timeframes that were tested.

For aggregate indicators, live validation is not limited to one “happy path” chart. The required matrix spans supported USD-like spot and perp charts, non-USD and index-like contexts, unsupported-base fallback contexts, Spot / Perps / Both market-type modes, and more than one timeframe. When relevant, it also expands into reset modes, anchor boundaries, cumulative behavior, and CVD-style source transitions.

Why This Matters to You

Most users will never read Pine code, and they should not have to. The point of the audit is to turn hidden technical risk into visible, repeatable quality control. It is how we reduce the chance of silent engine drift, bad-source handling, misleading debug output, unstable behavior across chart contexts, or documentation that no longer matches the script.

That matters because an indicator can look convincing and still be wrong in important ways. It might compile, plot, and even seem reasonable on one chart while still failing on source integrity, fallback honesty, reset correctness, or cross-context behavior. The 30-point audit exists to catch those failures before they become your problem.

What “Release-Ready” Means at CIS

For us, release-ready means more than “no syntax errors.” It means the indicator-specific math is correct, the shared engine is still in parity, the plotted source is the intended source, the debug layer is honest, the warmup and first-valid-bar behavior are truthful, the documentation matches the current script, and the final build has passed the release gate.

That is the standard behind the suite. Not just functional. Audited. Not just presentable. Accountable. And not just built to work once, but built to hold up across the chart conditions real users actually trade.