Narrative Cascade Planner
Map an approved brand narrative across every marketing surface and hand precise briefs to the teams that build them.
The skill treats pasted and scraped content as untrusted, limits registry changes to proposal events, and asks before saving results; however, sensitive-data handling, rollback after write failures, and the complete external-effect surface are not fully specified, so 7 points are deducted.
Triggers, inputs, outputs, completion criteria, and the NEEDS_INPUT stop path are coherent; however, the skill depends on several upstream skills, runtime scripts, and host capabilities, with no dedicated key-path tests or abnormal-input feedback evidence. Static calibration caps this at 10, with 2 points deducted.
The audience, use cases, non-fit tasks, argument hint, and English/Chinese descriptions are fairly clear; however, operation requires an existing canon, surface inventory, and multiple protocol artifacts, while mainland-China connector reachability and boundary evidence are limited, so 4 points are deducted.
The document has strong structure, Quick Start examples, a contract, data sources, save rules, references, versioning, and Apache-2.0 licensing; however, it lacks a skill-specific changelog, explicit maintenance owner, and update path, and SECURITY.md lists 17.x as supported while the skill is version 18.0.0, so 4 points are deducted.
The expected cascade plan, surface specifications, drift list, and handoff briefs are concrete and potentially directly usable by downstream builders; however, the result depends heavily on complete upstream inputs and there is no verified output example or execution evidence. Static calibration caps this at 7, with 1 point deducted.
The skill identifies the canon, claims ledger, baseline map, official store documentation, and related audit references, providing some traceability; however, the supplied files contain no skill-specific reproducible example, independent corroboration, or key-path test coverage. Static calibration caps this at 5, with 1 point deducted.
- The skill is not standalone: it stops without a sealed canon or surface inventory, so upstream artifacts must be supplied first.
- The skill declares version 18.0.0 while SECURITY.md lists 17.x as the supported line; confirm support and maintenance status before upgrading or installing.
- Although connector and scraping safeguards are documented, authorization, robots compliance, data egress, and mainland-China network reachability still require runtime verification.
- The output is a planning and handoff specification, not finished page, ad, email, or social copy; downstream builders and human review remain necessary.
What does this skill do, and when should you use it?
This skill maps an established narrative-registry canon onto the homepage, pricing page, store listing, sales deck, social bio, docs, and email. For each surface, it specifies the lead claim, narrative pillar, proof module, objection reframe, and localization note. It also produces handoff briefs for content, ad, email, social, and launch-message builders. It creates planning specifications rather than finished copy, claim adjudication, or cross-surface scoring.
Reads the narrative canon, the surface inventory and current-versus-intended gap, the read-only claims ledger, and available proof modules; creates a message-match specification for each flagship surface; identifies live copy that contradicts the canon's tagline, pillars, or approved wording; marks unsupported product or comparative claims as “needs source” and routes proposals; creates builder handoff briefs containing the canon pointer, surface specification, lead pillar, proof, angles, and voice constraints; and delivers the cascade plan, drift list, and briefs.
- A brand team has sealed its narrative canon and needs to apply it consistently across its homepage, pricing page, and documentation.
- A product team is preparing a launch and needs the store listing, sales deck, and email to use aligned claims and objection responses.
- A marketing lead wants to brief content, advertising, and email teams from one narrative source without having the planning skill write finished copy.
- A team has detected message drift between channels and needs each specific conflict named before publication.
What are this skill's strengths and limitations?
- Covers multiple owned marketing surfaces with explicit claim, pillar, proof, and objection requirements.
- Separates planning and handoff specifications from finished copy production.
- Routes unsupported claims for sourcing instead of presenting them as established facts.
- Uses the repository's shared skill contract and is compatible with Claude Code and compatible Agent Skills hosts.
- Requires an existing canon with a tagline, pillars, voice rules, and boilerplate; otherwise it stops for input.
- Requires a supplied surface inventory or baseline map and does not guess which surfaces exist.
- Does not write finished page, ad, email, or social copy, and does not perform the final consistency score.
- The supplied source does not include a real execution example or sample output.
How do you install this skill?
Install the repository with Claude Code using /plugin marketplace add aaron-he-zhu/aaron-marketing-skills, then /plugin install aaron-marketing@aaron; or use npx skills add aaron-he-zhu/aaron-marketing-skills on a compatible Agent Skills host, or clone it with git clone https://github.com/aaron-he-zhu/aaron-marketing-skills. The skill is at narrative/land/narrative-cascade-planner/SKILL.md.
How do you use this skill?
Provide memory/narrative-registry/canon.md, the narrative baseline map, the claims ledger, and the target surface list. Then trigger it with: Plan the narrative cascade for [product] from our canon. Surfaces: homepage, pricing, store listing, sales deck, social bio, docs, email. Another supported prompt is: Write a per-surface message-match spec for our pricing page against the canon and flag any drift. After delivery, it asks whether to save the results.