decorum — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited decorum (Agent Skill) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
Not what to say, but how to say it. Not tone, but tradition.
This skill emerged from a session that kept encountering register as the hidden variable in craft judgments. The Silmarillion's withholding of interiority wasn't a failure — it was fidelity to the Atlakviða tradition, where emotional display is beneath the poem's dignity. A flute-model's performative enthusiasm wasn't wrong in itself — it was the wrong register for close reading, which has its own decorum of precision and groundedness. Tolkien's tetrameter wasn't simple — it was hobbit-appropriate, deliberately written beneath the author's skill level to match the character's. In every case, the quality judgment was actually a register judgment: is this response operating in the right tradition of discourse for what it's trying to do?
Every task belongs to a tradition of discourse. That tradition has rules. The rules are the decorum. Violating the decorum undermines the content, even when the content is good.
A brilliant insight about Tolkien, delivered in the register of a LinkedIn post, is worse than a modest insight delivered in the register of close reading. The register isn't packaging. It's part of the meaning. How you say it shapes what you say, because the register carries implicit commitments about the relationship between speaker, listener, and subject.
Registers are not tones. Tone is surface (warm, cool, formal, casual). Register is deeper — it's the entire set of conventions that govern a tradition of discourse. Some registers this skill ecology works with:
Ask: What tradition of discourse does this task belong to? What are the rules of that tradition? If I violate those rules, what will happen to the content?
Check periodically: Am I still in the register I chose? Has the writing drifted into a different tradition? (This happens especially at transitions — the temptation to shift from close-reading register into exploratory register when a connection appears, or from diagnostic register into mockery when the diagnosis is damning.)
When something feels wrong — your own output or someone else's — ask: Is the register mismatched? Is this close reading written in an exploratory register? Is this mythic content delivered in a novelistic register? Is this diagnosis performed as entertainment? The mismatch is often the whole problem.
Sometimes a piece of writing deliberately mixes registers — Tolkien's Lord of the Rings alternates between novelistic warmth and mythic grandeur. This isn't a violation but a modulation, and the skill is in managing the transitions. Check: are the transitions motivated (by a change in scene, character, or stakes) or arbitrary? Motivated transitions are craft. Arbitrary transitions are mess.
Sometimes you don't know what register a task requires until you've started. This is fine. The skill isn't always pre-emptive — sometimes it's corrective. You start writing, realize the register is wrong, and adjust. The adjustment is the skill.
Rarely, a task requires a register that doesn't exist yet — a tradition of discourse that hasn't been formalized. This is where /decorum meets /meta-skill. If you find yourself needing a register you can't name, that might be a new skill wanting to be born.
The Silmarillion is cold because it should be cold. The Shire is warm because it should be warm. The diagnosis is precise because it must be precise. Know what tradition you're in. Follow its rules. The rules are not constraints — they're the conditions under which the content can do its work.
~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.