Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Review positioning and messaging across content
.claude/skills/davekilleen-messaging-audit/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -24% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 6% | 0% |
| case-11 | ✗→✓ | ▲ Improved | 88% | 0% |
| case-12 | ✗→✓ | ▲ Improved | 47% | 0% |
| case-13 | ✗→✓ | ▲ Improved | 78% | 0% |
Use this skill when multiple artifacts need comparison for positioning, value proposition, audience, terminology, or product claims and a canonical baseline can be cited. It also handles an audit that discovers no usable baseline, provided that limitation is explicit.
Do not use it to invent positioning, approve legal or factual claims, or silently rewrite brand copy. Not for publishing edits, changing the canonical baseline, or deciding that a variation is intentional without evidence and human authority.
date/time. If no authority or version is available, record “no canonical baseline” as an unknown rather than choosing the most polished artifact.
publication or update date, retrieval/as-of date, audience/channel, and scope.
and terms verbatim. Attach a source and date to each extracted item.
and do not treat similar words as equivalent when the source does not support that.
and as-of boundary.
conclusions while reporting them as coverage limits.
artifact without paraphrasing away meaningful qualifiers.
baseline and with other in-scope artifacts.
authority are evidenced, contradictory, stale copy, missing, or an unsupported claim. Do not call a difference intentional merely because it sounds appropriate.
state confidence, and surface unknowns and contradictions side by side.
or saved report and obtain explicit confirmation from the human authority.
artifact or canonical source.
wording; explain the inference and confidence.
boundary; do not call it a current contradiction without checking versions.
sources, dates, and any known scope difference.
An unsupported claim means “support was not found in the permitted sources”; it does not prove the claim false. Never invent dates, metrics, percentages, owners, intent, money, causes, status, or evidence. Recommendations are not human decisions.
Return an audit with:
or unsupported claim, each with source, date, evidence, and confidence;
approved messaging decisions.
Default to read-only. Do not rewrite, publish, approve, or replace canonical messaging; do not edit content systems or claim authority over legal, product, or brand decisions. A requested write requires a precise preview, explicit confirm, and human authority. Apply only that confirmed scope; a recommendation remains a recommendation.
Read back every finding against the cited artifact and reconcile wording, version, source, date, and classification before delivery. After an authorized edit or saved report, read back the destination and reconcile it with the confirmed preview and canonical baseline. If a baseline or artifact cannot be read, mark dependent findings unknown; if a newer version appears, mark prior comparisons stale and re-check. If a write fails or is partial, stop, preserve the draft and error, report what changed, and wait for human authority before retrying or recovering.
Other measured skills in the registry, with their headline benchmark lift.