What this catalogue can—and cannot—tell you.

The current evidence label

Conditional lifecycle/email candidate

Core Haines emails mutable selected text displayed 2.1.3 on 14 Oct. A 2.1.0 change commit is historical, not a verified latest file pin; the 2.1.1→2.1.3 diff and current inventory remain open. It is included only for our proposed five-email post-signup welcome Markdown hypothesis from an owner-approved, non-personal brief. The owner-maintained template is the first route for one local draft. Timing, exits and metrics are hypotheses, not send-ready rules or targets; a named owner checks proof, terms and the exact CTA URL. Tagged 2.1.4 adds reference-before-drafting and a local Learnings read versus complete tagged 2.1.0; these do not authorize access. Implementation URLs now point to mutable main. Do not read local product context without authorization. The linked Copywriting AI-tells page is cross-skill style guidance; the separate implementation guides were source-read on 14 Oct and describe credentials, recipient data and send routes. GitHub now lists an emails-source.zip release asset with reported SHA-256, and tagged v2.11.37 source displays 2.1.4. The ZIP bytes and attached manifest were inaccessible, so no entry is confirmed present or missing; CLI-selected payload, latest mutable file state and host/provider behavior remain unverified. Source reading gives no permission to implement. A human authorized for brand, recipient policy, and the email system must verify the draft and separately approve any import, configuration, scheduling, test, publication, or send. This is not an installation, compatibility, safety, privacy, deliverability, or outcome finding.

Read the 17 October release-asset boundary. Read the 11 October welcome-draft boundary. Read the earlier lifecycle/email candidate review.

Current stop boundary

Copywriting’s older selected-file record covers historical 2.0.2. A separate 2 October commit directly lists the selected file for the 2.1.0 change, not a verified latest main pin. Mutable selected text displayed 2.3.1 on 16 October; 2.2.0 was a dated 15 October observation. Full 2.1.0 and observed 2.3.1 endpoints were compared, but the exact complete 2.1.0→2.2.0 intermediate delta, latest pin, inventory, actual payload and host behavior remain unresolved. Run 22 reviewed the AI-tells reference only for style and evidence claims; it is not a validated detector. The source’s conversion figures and 3× example lack a checkable study in inspected passages. Current inventory still needs review. Recheck the source you intend to use. The 25 September recheck did not establish selected-file SHA/date or tag findings for the four Superamped entries examined then. The remaining two active Superamped cards retain that stop; Search Page Audit and Ad Campaign Analyzer are removed from active selection and are not recommended for installation. Stop also for credentials or account connection, paid use, sensitive customer or campaign data, untrusted web content, and any publication, send, deployment, or spend change without an authorized human approval.

Read the provenance recheck. · Read the evaluation and host-assumption report.

Why four remain active

The active set spans page-copy drafting, content planning, public-web research and one conditional lifecycle-email draft. Ad Campaign Analyzer was removed on 7 October; Search Page Audit was removed on 8 October after its score and claim audit. Paid-media exports and page audits have no recommended skill here. The selection favors reviewable, lower-access outputs over skills that guide production instrumentation, need paid credits, or enlarge the service and data surface.

Selection is not a ranking. RampStack email-sequences is excluded because its inspected workflow includes email-service-provider configuration, self-send, broadcast scheduling, and measurement routes. Deferred examples may be useful later, but need separate source inspection and a clear evidence record before entering the catalogue.

Runs 8–10: a bounded reporting search, now closed

The 28 September report proposes a later search for an offline, human-reviewed monthly cross-channel briefing skill using approved aggregate exports. Run 8 did not seek or inspect a package. Runs 9 and 10 each inspected two original reporting files; none requires the full fixed audit and owner-approval protocol. Run 10 closes this search for now. There was no workflow test, then-six-entry recheck, or seventh catalogue entry. Community anecdotes do not establish demand, time saved, or client outcomes. Read the Run 8 job. Read the Run 9 shortlist. Read the Run 10 closure.

Non-negotiable review boundary

1. Read the sourceInspect package files, install guidance, licenses, scripts, references, and assets.
2. Use minimum dataOnly provide material you are authorized to share; campaign and customer material may be sensitive.
3. Verify outputCheck facts, calculations, citations, brand fit, and assumptions.
4. Approve actionKeep publishing, sending, deployment, and spend changes behind a human approval.

Run 13: competitor-change brief skill search deferred

A source-paired review for one known rival can record observed public-page differences. Dated captures, comparable context and named human verification are required. A page difference does not prove strategy, rollout or impact. A saved diff, versioned sheet, reviewed prompt or monitor may suffice; no package or monitor was tested. Read the full Run 13 report.

Run 12: prelaunch QA skill search deferred

Run 12 defines an offline comparison of approved brief with dated proposed setup, and keeps cross-channel UTM governance separate. A discrepancy in copied records needs account-owner verification, not a readiness claim. No package was sought, installed or tested.

Read the full Run 12 report.

Run 11: launch-brief skill search deferred

The first-use brief is a local, owner-reviewed document for one approved offer and two named channels. Practitioner evidence shows brief-making, but not repeated failure of a template, reviewed prompt or connected campaign workspace. No package was sought or tested, and no seventh entry was added at that date. Read the decision and next QA question.

Run 26: Copywriting source-delta boundary

Observed 2.3.1 diagnoses a blank brief, real-voice draft or point edit. A point edit asks for changed lines and brief reasons, rather than default full-page copy and alternatives. Source claim-support directions do not replace the pool’s proof, rights, exact CTA, named rendered-page review or separate publication gates. The full 2.1.0→2.3.1 endpoint comparison does not settle the exact intermediate 2.1.0→2.2.0 delta or an installed payload. Read Run 26.

Run 22: Copywriting claim/proof boundary

For one approved local landing-page draft, use an owner-maintained brief/template first. Copywriting supplies structure, not verified conversions or AI detection. The pool-proposed claim register holds unsupported numbers, testimonials, comparisons, guarantees and CTA links; a named owner checks facts and the rendered page before a separate publication decision. Selected-file pin, inventory, installation and quality remain unverified. Read Run 22.

Run 21: welcome-email draft boundary

The mutable emails selected text displayed 2.1.3 on 11 October; latest file pin and full inventory remain unknown. Our five-email post-signup welcome brief is only a local draft hypothesis. A named owner reviews claims, proof, terms and exact CTA URL; recipient-policy and ESP owners approve rules and later action separately. Neither template nor skill was tested. Read Run 21.

Run 20: Competitor Discovery classification and ranking boundary

Retain only a human-confirmed, dated public-web starting set. Direct and adjacent types are provisional; the source normally excludes tangential results. It ranks candidates by search frequency, but this is not verified threat or action priority. A versioned owner sheet and reviewed search come first; dated row evidence and separate approval are needed before claims or actions. No selected-file pin, current inventory or installation advice. Read Run 20.

Run 19: Content Strategy priority boundary

Content Strategy remains only a conditional draft-planning scaffold. Its X/10 weights and gap or buyer-stage labels are hypotheses, not measured opportunity. For one approved first-month plan, use an owner-maintained versioned brief/template first. Selected-file provenance and current inventory remain open; no install recommendation. Read Run 19.

Run 18: search-page score boundary

Search Page Audit is not currently recommended. Its mandatory X/38 score, ratings and fixed priorities are not validated search-quality, AI-citation or ranking measures; several checks conflict with current Google guidance. For one authorized public URL, use the owner-maintained, unscored checklist and named SEO/site reviewer. Its selected-file SHA/date and current inventory remain unverified; no install or deployment. Read Run 18.

Run 17: paid-media decision boundary

The selected Ad Campaign Analyzer text uses AOV minus ad CPA as profit and ten conversions as a significance cue while directing immediate stop/scale action. These are source rules, not validated thresholds. The active card was removed. Use a governed descriptive worksheet and named paid-media owner; no account or spend action is authorized. Selected-file SHA/date, tag and current inventory remain unresolved. Read the audit.

Run 16: changed-source check

Copywriting and emails have file-attributable 2.1.0 change commits, not verified latest file pins. Emails displayed 2.1.1, but its file SHA/date and full diff remain unknown. Read Run 16.

Run 15: route-first chooser

Consider a brief/template and reviewed prompt first for one local artifact; host-native instructions for repeated private context; a portable third-party file only after documented omissions and full package review; or a governed connected product for authorized live operations. These are untested route choices, not a ranking. The historical Search Page Audit row is superseded by Run 18. Read the historical six-entry decision/access/review matrix and its Run 17 and Run 18 superseding notes.