Narrative Baseline Mapper
Inventory what every owned brand surface says now, expose gaps against intended messaging, and freeze a baseline for future narrative-drift checks.
The skill explicitly treats pasted and fetched content as untrusted data, limits inventorying to owned surfaces, requires robots preflight, and places confirmation and authorized proposal gates around writes. However, delegated fetching can expose URLs and content to third parties, while runtime permission boundaries, rollback, and sensitive-data handling are not fully specified in this skill, so points are deducted.
The Skill Contract, done criteria, termination rules, source labels, ambiguity stop, and failure-oriented handling are relatively clear. However, this was a static review: connector behavior, write paths, failure messages, and key-path reproduction were not executed or proven by this file; the static ceiling therefore applies.
Triggers, inputs, outputs, non-fit boundaries, and downstream routing are clearly stated, and pasted data provides a tool-independent fallback. However, the core scraping path may depend on external services, Chinese semantic quality is not demonstrated, and mainland-China reachability is not specifically established, so points are deducted.
The documentation is well layered and includes versioning, licensing, argument hints, quick starts, data sources, limitations, save confirmation, and next-skill guidance. Maintenance ownership and update procedures are not clear; the repository security policy references supported 17.x while this skill declares 18.0.0; and publisher identity is unverified, so points are deducted.
The task, per-surface inventory, evidence labels, gap categories, baseline snapshot, and handoff format are concrete and potentially directly usable. However, no executed key-path evidence or representative verified output is supplied; connector access, memory writes, and claim proposals still require environment support and review, so the static ceiling and uncertainty reduce the score.
The skill requires source attribution, as-of dates, Measured/User-provided/Estimated labels, and exact quoted evidence, providing some auditability. There is no independently reproducible result, third-party corroboration, or skill-specific execution evidence; the supplied repository tests mainly cover shared architecture and artifact validation, so points are deducted.
- This is a static review only; the skill, connectors, and write paths were not executed, so documented safety and usability claims remain unverified.
- Before using delegated fetching such as Firecrawl, confirm that the target may be shared with the provider and that the user owns or is authorized to access it; also verify reachability and robots rules.
- The repository security policy and the skill declare different version lines; confirm maintenance status and compatibility before deployment.
What does this skill do, and when should you use it?
Narrative Baseline Mapper is the Trace-phase skill for the repository’s Narrative discipline. It records the current wording on owned surfaces such as homepages, pricing pages, docs, decks, social bios, and email signatures. Each item is labeled Measured, User-provided, or Estimated, then classified as aligned, drifted, contradictory, or silent against an intended message or existing canon. It also flags unsupported claims and freezes a dated, sourced snapshot for later drift monitoring. It does not author the canon, score surfaces, or run TALE vetoes.
Reads user-pasted copy, owned-surface material, optional Wayback history, and any existing narrative canon in project memory; extracts current headlines, value lines, one-liners, and claims by surface; labels evidence as Measured, User-provided, or Estimated; compares each surface with the stated intended message or existing canon; records exact quotes supporting each gap classification; flags unverifiable product or comparative claims as [needs source]; and produces a surface inventory, gap read, frozen drift baseline, and handoff summary. It asks before saving results to project memory.
- A repositioning team needs an evidence-based inventory of what its homepage, pricing page, and docs currently communicate.
- A marketing lead suspects inconsistent value propositions across channels and wants the exact wording behind each drift or contradiction.
- A product team wants to freeze a pre-repositioning snapshot of every owned surface.
- A brand has an approved one-liner and wants to check each owned touchpoint against it.
- A team needs to collect unsupported product claims for later substantiation by the claims registry.
What are this skill's strengths and limitations?
- Has a narrow, explicit scope: current-state inventory, gap reading, and drift-baseline capture.
- Preserves exact surface wording with source and as-of date for reviewability.
- Works at Tier 1 with pasted data or keyless retrieval; paid tools are not required.
- Separates measured evidence, user-provided evidence, and estimates.
- Provides a defined handoff to downstream narrative skills such as category-narrative-mapper and narrative-drift-monitor.
- Does not write the brand canon, message house, or final copy.
- Does not score surfaces or calculate TALE results, and does not adjudicate claims.
- Closed-platform social bios must be supplied as pasted copy rather than scraped.
- Retrieval, history lookup, and memory writes depend on the repository’s referenced scripts and authorized project-memory workflows.
- The supplied source does not document an independent test suite or platform-coverage report for this individual skill.
How do you install this skill?
Install the collection in Claude Code with /plugin marketplace add aaron-he-zhu/aaron-marketing-skills followed by /plugin install aaron-marketing@aaron. On a compatible Agent Skills host, run npx skills add aaron-he-zhu/aaron-marketing-skills, or clone it with git clone https://github.com/aaron-he-zhu/aaron-marketing-skills. The skill file is narrative/trace/narrative-baseline-mapper/SKILL.md.
How do you use this skill?
Provide a brand or product name, URLs or pasted copy for the owned surfaces, and an intended message if one exists. Example: “Map what Acme’s homepage, pricing page, docs, and social bios say today, and show the gaps against ‘help small teams close their books faster.’” If neither an intended message nor an existing canon exists, the skill records that no canon is available instead of inventing a yardstick.
How does this skill compare with similar options?
Unlike message-system-architect, this skill inventories existing messaging rather than authoring a canon or message house. Unlike narrative-quality-auditor, it does not score surfaces or run TALE vetoes. Unlike category-narrative-mapper, it focuses on owned surfaces rather than category and competitive narratives. Unlike positioning-truth-tracer, it does not reconcile positioning claims against shippable reality and substantiation.