Brand Message System Architect
Build a durable brand hierarchy connecting the narrative, value pillars, proof points, tagline, and one-liner.
The skill treats pasted material as untrusted, labels unsupported claims as [needs source], uses an authorized propose flow for registry writes, and asks for confirmation before saving. Repository security material also describes no third-party dependencies, no credential persistence, and locked, atomic registry operations. Seven points are deducted because the skill does not fully specify least-privilege permissions, sensitive-data handling, failure rollback, or host capability configuration, while proposals still create external state changes.
The instructions define upstream-input checks, NEEDS_INPUT, scope guards, consistency checks, evidence labels, and termination rules, and are broadly self-consistent. Twelve points are deducted because static files do not show dedicated tests for this skill, reproduced key paths, robust abnormal-input handling, or diagnosable failure messages; static calibration therefore limits the score to 10 or below.
The trigger, target scenario, output scope, and non-fit boundaries relative to message-house-builder, offer-claims-registry, and strategic-narrative-designer are clear. It does not require overseas connectors and includes Chinese naming and descriptions. Four points are deducted because it depends on specific upstream artifacts, memory paths, registry capabilities, and host behavior, while evidence for precise semantic routing and avoidance of false triggers is limited.
The skill provides structured metadata, Quick Start, Skill Contract, Data Sources, Instructions, Save Results, references, handoffs, a version, and an Apache-2.0 license. Five points are deducted because maintenance ownership, update path, changelog, and installation details are mostly repository-level rather than skill-local, and the security policy's supported 17.x line conflicts with the skill's 18.0.0 version.
Expected outputs, done criteria, proof-point labels, claim handling, consistency checks, and handoff artifacts are concrete and plausibly cover the core brand-message task. Nine points are deducted because no representative executed output or execution evidence is supplied, so correctness, completeness, and marginal benefit over manual work cannot be verified statically; static calibration caps this at 7.
The skill requires traceability to positioning truth, the claims ledger, prior canon, and claim IDs, and separates Measured, User-provided, and [needs source] evidence. The repository also supplies some general CI tests. Six points are deducted because there are no dedicated tests or independent corroboration for this skill's key paths, and the referenced source contents and registry submission behavior are not fully included; static calibration caps this at 5.
- The skill submits narrative and claim proposals; before execution, show the target paths, proposed changes, and confirmation points explicitly.
- Positioning truth, the narrative arc, personas, and the claims ledger must be supplied and checked first; otherwise return NEEDS_INPUT rather than inventing positioning or proof.
- Resolve the governance mismatch between skill version 18.0.0 and the security policy's supported 17.x line, and verify that the host supports the authorized registry-events.py flow.
- Measured and User-provided labels are not independent validation; unsupported product or comparative claims still require human review.
What does this skill do, and when should you use it?
Message System Architect creates a brand-level messaging system designed to outlast any single launch. It produces the main narrative, exactly three value pillars linked to named positioning themes, persona-specific proof points with evidence labels, and the tagline or one-liner. It also runs an internal consistency pass and prepares candidate submissions for the narrative and claims registries. It does not build a launch-specific message house, design the narrative arc, codify brand language, adjudicate claims, or score TALE.
Reads the positioning truth set, change-narrative arc, user-provided personas, the read-only claims ledger, and any existing narrative canon; verifies upstream inputs and prior canon; writes the main narrative, three value pillars, persona proof points, tagline, and one-liner; labels proof points as Measured, User-provided, or [needs source]; checks consistency across the hierarchy; identifies unapproved claims; submits claim and narrative candidates through authorized registry-events.py propose requests; and writes the working house to the designated memory directory.
- A brand team has a positioning truth set and narrative arc and needs a reusable, brand-level message house.
- A marketing lead needs three traceable value pillars and the proof each persona would require.
- A brand is preparing to submit a new narrative, tagline, and pillar set to its canon and needs contradiction checks first.
- A team needs to identify unsupported product or comparative claims before registry review.
What are this skill's strengths and limitations?
- Produces a concrete hierarchy covering the main narrative, three value pillars, persona proof points, tagline, and one-liner.
- Requires every pillar to trace to a named positioning value theme, preventing orphan pillars.
- Explicitly labels Measured, User-provided, and [needs source] evidence states.
- Fits a defined handoff path into the narrative registry, claims registry, and brand-language codifier.
- The repository is Apache-2.0 licensed and supports Claude Code and compatible Agent Skills hosts.
- Requires a complete positioning truth set and change-narrative arc; it does not invent missing positioning.
- Does not substantiate or adjudicate claims; unsupported claims require downstream handling.
- Does not write the formal narrative canon directly; it submits a candidate event.
- The persistence workflow requires project-memory filesystem access and authorized registry-events.py propose requests.
- The provided source does not include independent test results or a separate maintenance record for this skill.
How do you install this skill?
Install the collection with npx skills add aaron-he-zhu/aaron-marketing-skills, or clone https://github.com/aaron-he-zhu/aaron-marketing-skills. The target skill is narrative/architect/message-system-architect/SKILL.md. The repository also documents Claude Code installation: /plugin marketplace add aaron-he-zhu/aaron-marketing-skills, followed by /plugin install aaron-marketing@aaron.
How do you use this skill?
Provide the brand, personas, positioning truth set, and narrative arc. Example: Author the durable brand message house for [brand] from the positioning truth set and narrative arc. Personas: [list]. If either upstream artifact is missing or incomplete, the skill stops with NEEDS_INPUT. After delivery, results are saved only after confirmation.