Scores narratives, positioning blurbs, GTM briefs, and campaign summaries across 7 brand calibration dimensions before they drive external campaigns or circulate beyond the immediate team. Invoke when any directionally relevant narrative needs a framing and calibration pass — not a relevance filter, not a line-edit, not a content quality review. Produces a scored verdict (pass / warn / fail) with flagged bullets, specific paragraph references, and downstream impact notes. Distinct from content-review (which scores publication-ready content pieces across 8 quality dimensions) and positioning-review (which gates positioning statements before Layer 3 copy is written).
Installer avec Codex ou Claude Copiez ce prompt, collez-le dans Codex, Claude ou un autre assistant, puis laissez-le vérifier la page du skill et l'installer pour vous.
Une commande directe contourne le prompt de vérification. Examinez la source avant de l'exécuter.
Scores narratives, positioning blurbs, GTM briefs, and campaign summaries across 7 brand calibration dimensions before they drive external campaigns or circulate beyond the immediate team. Invoke when any directionally relevant narrative needs a framing and calibration pass — not a relevance filter, not a line-edit, not a content quality review. Produces a scored verdict (pass / warn / fail) with flagged bullets, specific paragraph references, and downstream impact notes. Distinct from content-review (which scores publication-ready content pieces across 8 quality dimensions) and positioning-review (which gates positioning statements before Layer 3 copy is written).
triggers
["/brand-calibration","/calibrate-narrative","user wants to check narrative before it goes external","user says \"does this messaging feel right\"","user says \"check the tone and claims on this\"","GTM brief produced by /gtm-brief before circulation","competitive intelligence summary before team distribution"]
Do not stack this review on top of content-review for the same artifact. This skill checks framing, calibration, and consistency. Content-review checks publication readiness. Run brand-calibration on strategy docs and briefs; run content-review on content pieces.
Contract
This skill guarantees:
All 7 dimensions are scored before verdict is rendered
Every flagged item includes a specific paragraph or section reference — no vague "the tone feels off"
PASS requires all 7 dimensions at 0.7 or above
WARN requires average ≥ 0.6 with no dimension below 0.5
FAIL triggered by any dimension below 0.5, OR any critical structural problem (strawman competitor, unsubstantiated absolute claim, contradictory messaging threads)
Downstream impact is stated for every flagged dimension
This skill does not re-litigate raw fact-gathering unless a claim obviously contradicts cited evidence
Brain reads / writes
Before starting, if a companion aether-growth-brain repo is connected:
Read knowledge/icp-map.md — tone match scoring requires a validated ICP definition; score against the named ICP's context, not a generic "technical audience"
Read knowledge/competitor-map.md — competitive framing scoring requires the approved competitive positioning; flag deviations from named alternatives and framing decisions
Read playbooks/messaging.md — messaging consistency scoring requires the approved messaging hierarchy; flag any claims, category terms, or value attributes that deviate
On completion, if verdict is WARN or FAIL, write a one-line flag to decisions/brand-calibration-flags.md:
[date] [artifact-name] | verdict: [WARN/FAIL] | top flag: [dimension — specific issue]
Inputs
Required before proceeding:
The artifact to calibrate: narrative, blurb, brief, competitive summary, or messaging doc. Paste full text or provide a link.
Intended audience (ICP) if not embedded in the artifact. If ICP is unspecified, calibration will note this and score tone match conservatively.
If no artifact is provided:
NEEDS_CONTEXT. Return:
"Please provide the artifact to calibrate. This can be:
- A narrative or positioning blurb
- A GTM brief or strategy summary
- A competitive intelligence summary
- Any messaging doc before external circulation
Paste the full text."
Dimension scoring guide (0.0–1.0 each)
Dimension 1 — Tone match
Does the voice fit the ICP and category context?
Score
Description
0.9–1.0
Voice is precisely calibrated for the stated ICP: precise and evidence-grounded for senior engineers; ROI-grounded for enterprise buyers. Leads with what the reader can stop doing, not only what they can start. No hollow hype; no unwarranted hedging that kills credibility.
0.7–0.8
Mostly right; one section tips into marketing adjectives or unwarranted hedging
0.5–0.6
Mixed — practitioner in parts, marketing in others; inconsistent register across sections
0.3–0.4
Primarily marketing voice; benefit language dominates over mechanism language
0.0–0.2
Generic pitch; no ICP signal; hollow enthusiasm openers; reads the same for any product in the category
Friction-elimination check: Does the narrative lead with what the reader can stop doing as a result of using this? Friction-elimination framing lands faster with practitioners living with the friction. If the piece leads only with what the reader can start doing, flag for reframe consideration.
Hedging check: Flag both directions — hype AND unwarranted hedging. "We think this might potentially be useful for some teams" from a product with clear, provable value is as miscalibrated as "game-changing."
Dimension 2 — Claim strength
Are claims calibrated to their evidence base?
Score
Description
0.9–1.0
All high-confidence claims have multi-source support in or attached to the doc; single-source claims are labeled "likely" or "based on [source]"; no overclaim or underclaim; all proof attributed to named company + named role + specific metric
0.7–0.8
One minor claim at higher confidence than its evidence; no absolute claims without proof
0.5–0.6
One or two claims asserting "high confidence" with only single-source support; or one claim with anonymous attribution
0.3–0.4
Multiple overclaims; or a core claim (the primary value assertion) is unsubstantiated
0.0–0.2
Absolute claims without any evidence; anonymous customer references throughout; "best-in-class" or "market leader" without proof
Overclaim flags: "best", "only", "always", "never", "everyone", "the future of X", "industry-leading" without a specific benchmark, date, and methodology.
Underclaim flags: Valuable, provable differentiators hedged away with "might", "could potentially", "may help" where "does" and "reduces X by Y" would be accurate and more useful.
Proof attribution rule: Named company + named role + specific metric = signal. "A large enterprise customer" = noise. "A Fortune 500 company" = noise. "The platform lead at a logistics company processing 400M events/day" = acceptable under NDA.
Dimension 3 — Competitive framing
Are named competitors handled factually, fairly, and in alignment with stated positioning?
Score
Description
0.9–1.0
Competitors named factually; no strawmen; no unnecessary elevation; framing critiques the architectural assumption or the era, not the vendor; consistent with the competitive positioning map
0.7–0.8
Mostly clean; one comparison that is technically accurate but framed slightly uncharitably
0.5–0.6
One comparison the vendor would dispute; or a competitive advantage stated without evidence
0.3–0.4
Strawman framing; competitor described at their weakest rather than their realistic positioning
0.0–0.2
Named competitor attacked rather than analysed; legally and reputationally exposed; or contradicts the approved competitive map
Preferred pattern: Critique the architectural assumption or the era, not the named competitor.
Preferred: "[Legacy approach] was designed for [old constraint] — that constraint no longer holds because [specific reason]."
Avoid: "[Competitor] doesn't support [feature]" — often disputed, easily dated, and invites counter-attacks.
Elevation check: Unnecessarily elevating a smaller competitor by naming them alongside leaders gives them unearned positioning. Name only when the comparison is fair and strategically necessary.
Dimension 4 — Messaging consistency
Do the narrative threads hold together without contradiction?
Score
Description
0.9–1.0
Market context ↔ competitive moves ↔ recommended actions all align; no contradictions; if a tension exists, it is named and explained
0.7–0.8
Mostly consistent; one section uses slightly different framing than the rest
0.5–0.6
Two threads in tension; reader must resolve the contradiction themselves
0.3–0.4
Contradictory claims in different sections of the same doc; market framing contradicts recommended motion
0.0–0.2
Core contradiction — e.g. the positioning claims one ICP while the recommended actions target a different one
Consistency check axes:
ICP named in the positioning ↔ ICP assumed in the recommended actions
Market context described ↔ urgency framing of the brief
Competitive alternative named ↔ value attributes claimed as differentiators
Category terms used in opening ↔ category terms used in closing
Dimension 5 — Category-language investment
Does the narrative deliberately build on category terms the product owns or wants to own?
Score
Description
0.9–1.0
1–2 category terms used consistently throughout; terms are reinforced, not introduced once and abandoned; no new category terms introduced inconsistently
0.7–0.8
Category terms present but used inconsistently (same concept described differently in different sections)
0.5–0.6
No deliberate category term investment; the piece uses the competitor's category framing by default
0.3–0.4
Category language actively works against the positioning — e.g. uses a legacy category name that positions the product as a subset of an incumbent
0.0–0.2
No category awareness; generic category language throughout; could be re-labeled as any competitor
Category term audit: List the 1–2 terms the product owns or wants to own. Flag if the artifact (a) uses these 0 times, (b) introduces them inconsistently with the approved definition, or (c) introduces a competing category term not in the approved messaging.
Dimension 6 — Friction-elimination framing
Does the narrative lead with what the reader can stop doing, not only with what they can start?
Score
Description
0.9–1.0
Opens with or prominently features the friction being eliminated; "stop doing X" or "no more Y" is explicit; reader identifies their current friction immediately
0.7–0.8
Friction-elimination present but secondary to capability announcement; reader has to infer it
0.5–0.6
Benefit-led rather than friction-led; reader sees the outcome but not the pain being removed
0.3–0.4
Feature-announcement framing throughout; no friction named
0.0–0.2
Pure capability list; reader must self-translate to their own pain
Why this dimension: Practitioners living with friction respond to friction-elimination framing faster than capability framing. "You can stop doing X" is a shorter path to recognition than "You can now do Y." This dimension checks whether the narrative takes that path.
Dimension 7 — Philosophical critique check
When the narrative makes competitive or category arguments, does it critique the structural assumption or the era — not the named vendor?
Score
Description
0.9–1.0
All structural critiques target the architectural assumption, the design era, or the constraint that has changed — never the vendor; critique is forward-looking and analytically grounded
0.7–0.8
Mostly structural; one instance where a vendor is named in a critique that would be cleaner as a structural argument
0.5–0.6
Mix of structural and vendor-targeted critique; the vendor-targeted instances are accurate but could be reframed
0.3–0.4
Vendor-targeted critique is the primary frame; structural argument is secondary or absent
0.0–0.2
Attack-mode language; no structural argument; legally exposed; trust-destroying with the technical audience who knows the competitor
Structural critique template:
"Tools designed for [old constraint] carry [specific structural limitation] — not because of poor engineering, but because that constraint no longer applies."
"The [legacy approach] era assumed [specific assumption]. That assumption no longer holds because [specific evidence]."
Verdict logic
PASS: All 7 dimensions ≥ 0.7
"Ready for circulation. [Optional: one improvement to consider.]"
WARN: Average ≥ 0.6 AND no dimension < 0.5
"Circulate with noted cautions. Fix flagged items before driving campaigns."
Provide specific fixes for every dimension < 0.7.
FAIL: Any dimension < 0.5
OR any of: strawman competitor, unsubstantiated absolute claim, contradictory messaging threads
"Do not circulate. Fundamental calibration problem."
Name the structural issue and the right reframe.
Output format
## Brand Calibration Review
**Artifact:** [title or description]
**Type:** [narrative / blurb / GTM brief / competitive summary / messaging doc]
**ICP context:** [named ICP or "not specified — scored conservatively"]
**Reviewer:** brand-calibration
**Date:** [date]
---
### Dimension scores
| # | Dimension | Score | Note |
|---|---|---|---|
| 1 | Tone match | [0.0–1.0] | [one line] |
| 2 | Claim strength | [0.0–1.0] | [one line] |
| 3 | Competitive framing | [0.0–1.0] | [one line] |
| 4 | Messaging consistency | [0.0–1.0] | [one line] |
| 5 | Category-language investment | [0.0–1.0] | [one line] |
| 6 | Friction-elimination framing | [0.0–1.0] | [one line] |
| 7 | Philosophical critique | [0.0–1.0] | [one line] |
**Average:** [calculated]
---
### Verdict: [PASS / WARN / FAIL]
---
### Flagged items
[Each flag: dimension name, paragraph/section reference, specific issue, specific fix]
**Flag 1 — [Dimension]**
- Location: [paragraph number or section heading]
- Issue: [specific problem — quote the phrase if short]
- Fix: [what to change and how — specific, not "make it more specific"]
- Downstream impact: [what goes wrong if this isn't fixed before campaigns run]
**Flag 2 — [Dimension]**
- Location: [paragraph number or section heading]
- Issue: [specific problem]
- Fix: [specific rewrite or reframe]
- Downstream impact: [specific consequence]
---
### Downstream weighting notes
[What downstream work should weight more or less based on this calibration. E.g.: "If this brief drives paid social, the claim strength issues must be resolved first — unsubstantiated claims in paid copy carry legal exposure and damage trust with the developer audience."]
---
### What scored well
[Patterns to preserve and repeat. Do not omit this section — calibration is not only critical.]
Anti-patterns (reviewer anti-patterns)
Anti-pattern
Why it fails
Fix
Vague flags ("the tone feels off")
Author cannot act on it
Every flag must cite a specific paragraph or sentence and a specific rewrite direction
Ignoring underclaim in favor of overclaim hunting
A product with real, provable value that hedges every claim loses the reader's attention
Score underclaim as a calibration failure, same as overclaim
Treating all competitor mentions as a problem
Accurate, structural competitor comparisons are strategically necessary and legally safer than no comparison
Score the framing (structural vs vendor-attack), not the presence of competitor names
Giving WARN when a dimension scores 0.4
Average-washing hides structural failures
Any dimension < 0.5 triggers FAIL; do not average it away
Skipping downstream impact
Flagged items without downstream impact can feel cosmetic; owner deprioritizes them
Every flag must state what goes wrong if it isn't fixed before campaigns run
Re-litigating fact-gathering
This skill checks framing, not primary research
Do not question whether the underlying facts are correct unless a claim obviously contradicts the cited evidence already in the document
Scoring without ICP context
Tone match requires a specific ICP; "technical audience" is not enough
Ask for ICP context if not embedded; score conservatively and note the assumption
Validation criteria
Before delivering the review:
All 7 dimensions scored
Average calculated
One of three verdicts applied using verdict logic
Every dimension scoring < 0.7 has a specific flag with paragraph reference and fix
Downstream impact stated for each flag
"What scored well" section populated
If WARN or FAIL and brain is connected: flag written to decisions/brand-calibration-flags.md
References & Sources
Tier 1 (authoritative frameworks):
April Dunford, Obviously Awesome (2019) — competitive alternative framing, positioning consistency logic
Tier 2 (operator templates — adapted, not authoritative):
brand-narrative-calibration (growth-skills v1.0, score 6.5/10): original 4 dimensions (tone match, claim strength, competitive framing, messaging consistency), output shape, "do not re-litigate raw fact-gathering" rule
