Skip to content

hashgraph-online/awesome-codex-plugins / plugins/alpoxdev/hypercore/skills/hyper-google-tagmanager

hyper-google-tagmanager

Verified

Use this skill when the user asks to install Google Tag Manager on a website, audit an existing GTM install or a Google tag that runs beside a container, design or change a GTM container (dataLayer, tags, triggers, variables, consent, ecommerce), verify that tags fire, diagnose broken or duplicate tracking, or analyze what a container and its data show. It reads the repository, the served page, and a container export with bundled scripts, writes framework-specific install code, gates container publishing and live consent changes behind explicit approval, and saves a report under `.hyper/gtm/`. Do not use for GA4 property administration or reporting alone, ad campaign setup, or a tracking plan written as a document.

rules/validation.md

View on GitHub
# Validation

**Purpose**: Define the evidence each mode must have before the run reports success, so a working tag is never claimed from a status code, a dashboard glance, or the agent's own reading of the code.

## 1. Completion Evidence Per Mode

| Mode | Evidence required |
|---|---|
| `audit` | The JSON of each script that ran; for every finding, the file and line, the page offset, or the container item it came from; every page listed with the method used (served HTML or rendered browser) |
| `install` | The diff of the changed files; a second `gtm-detect.sh` run with one container ID, one loader, and no direct Google tag for the same ID; the `verify` table |
| `configure` | The workspace items or import JSON that changed; a `gtm-container-audit.sh` run after the change with no new `error` finding; the `verify` table; the publish result (version ID) or the proposed publish action |
| `verify` | One row per planned tag, as in §2 |
| `diagnose` | The symptom, the layer that explained it, the change, and the same check before and after |
| `analyze` | The container counts and findings, the plan-versus-container comparison, and for data checks the report, date range, and data-quality flags |

## 2. Verification Row

| Column | Content |
|---|---|
| Tag | The tag name and type |
| Trigger event | The dataLayer event or built-in event that fired it |
| Fired | `yes`, `no`, or `blocked by consent` from Preview, Tag Assistant, or the rendered page |
| Request | The destination request (for example a `/g/collect` request with its `en` value) or `none` |
| Key parameters | The parameters the plan requires, as seen in the request or the debug view |
| Consent state | The consent types and their state at the time |
| Result | `verified`, `failed`, or `not verified` with the reason |

A tag is `verified` only when its request or debug event was observed in this run. When no browser or debug capability exists, mark the row `not verified` and give the exact check the user can run.

## 3. Disallowed Proof

- `gtm.js` answering HTTP 200. An unknown container ID also answers 200 with an empty container (resource version `1`, no tags).
- A GA4 standard report from the last 24 to 48 hours, or the absence of data there.
- The code reading correct, or a script finding list that is empty.
- A Tag Assistant run with "no tags found" when ad blockers, CSP, an iframe, or a late-loading tag can explain it.

## 4. Report Gate

- `.hyper/gtm/<slug>/report.md` exists and follows `references/report-template.md`.
- Every number in the report has its source (script output, request, or report) and its date.
- Approval-tier actions appear either with the approval quote and their result, or in the proposed actions list.
- The residual-risk section lists every `not verified` row and every finding left open.
- The `결과:` line names exactly one of the five values in §5 and agrees with what ran.

## 5. Terminal Result

The report's `결과:` line is the value the next stage reads, so it names one outcome, not a summary. Pick it in this order:

| Value | When |
|---|---|
| `blocked` | The run cannot continue without something only the user can supply and an approval would not unblock it: a missing container ID, no site or repository access, a container export that was never provided, a credential the API step needs |
| `exhausted` | A `diagnose` root cause spent its three attempts without a passing proof, or a repair would have to be reclassified under a new name to continue |
| `proposed` | The work stopped at the approval gate with at least one approval-tier action listed: a publish, a republish, an overwrite import, a deletion, a permission change, or a live consent change |
| `partial` | The mode's required evidence is incomplete for a reason that is not approval and not the diagnosis budget: a `not verified` row left by a missing browser or debug capability, a page that could not be reached, a check the run could not run |
| `verified` | Every check the mode requires ran and each claimed tag has its request or debug event: a proven `install`, a proven `verify`, an `audit` or `analyze` whose findings all carry evidence and whose mode made no edits |

- `verified` is never the default. A run that published after approval and re-verified on the live page is `verified`; the same run that published and could not re-verify is `partial`.
- An `audit` or `analyze` whose requested scope all ran and whose findings all carry evidence is `verified`; one with a check that could not run is `partial`, and the report's residual-risk section names that check.
- The value describes this run's evidence, not the container's health. A container full of `error` findings still ends `verified` when the audit ran and every finding is evidenced.

## Sources

> Checked 2026-11-08.

| Claim | Source |
|---|---|
| Standard reports take 24 to 48 hours; network requests and DebugView are the immediate checks | <https://developers.google.com/analytics/devguides/collection/ga4/troubleshoot> |
| Tag Assistant connection failures from ad blockers, CSP, iframes, late loading, and AMP | <https://support.google.com/tagassistant/answer/10039345> |
| Unknown container IDs answer 200 with an empty container | execution check, 2026-11-08, `references/verification.md` §4 |

### Evidence grade

`LOCAL` - the evidence table is this package's design over the sources above.
rea-analysis-optimizationmorluto/reaOptimize REA providers and shared analysis workflows for agent context, latency, CPU, and memory while preserving correctness and Evidence. Use for analysis bottlenecks and cache or query design; CI/test-suite optimization belongs to rea-ci-tests-audit.Stars: 467kUpdated Claude CodeCodexVerifiedInstall
rea-ci-tests-auditmorluto/reaAudit REA CI and test suites for correctness, fixture drift, false passes, and measured performance; use for suite or workflow investigations and efficiency changes, rather than ordinary feature testing or a single test failure.Stars: 467kUpdated Claude CodeCodexVerifiedInstall
rea-tool-designmorluto/reaAudit, investigate, design, or change REA investigation tools, CLI/MCP contracts, provider capabilities, and Evidence semantics.Stars: 467kUpdated Claude CodeCodexVerifiedInstall
ponytail-auditDietrichGebert/ponytailQuality audit of the whole repo: bugs, security, real load, missing tests, speed, and what to delete. Most important first.Stars: 193kUpdated Claude CodeCodexVerifiedInstall