ClawHub

Brand Voice Architect

v0.1.1

A high-precision engine for deconstructing, documenting, and synthesizing brand-specific linguistic patterns and tonal architectures. Use this skill whenever...

⭐ 0· 116·0 current·0 all-time
byDavid Escobar@midnightstudioai
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
VirusTotalVirusTotal
Benign
View report →
OpenClawOpenClaw
Benign
high confidence
✓
Purpose & Capability
Name/description match the included artifacts: a corpus analyzer and a prompt synthesizer. Required resources (none) align with an instruction-only text-processing skill. There are no unrelated credential, binary, or config requirements.
ℹ
Instruction Scope
SKILL.md stays within the stated purpose (analyze corpus, synthesize voice prompts, produce guides). One notable behavior: generated system prompts instruct models to 'replace prohibited words with preferred equivalents silently' and to 'Never break voice for clarity alone' — this can cause automated alterations to user content and potential meaning drift. The workflow does include a human review step, which mitigates this risk if followed.
✓
Install Mechanism
No install spec or external downloads. The skill ships two small Python scripts and a methodology doc. No network fetches, URL installs, or archive extraction are present.
✓
Credentials
The skill requests no environment variables, credentials, or config paths. The scripts only read user-supplied corpus files or inline text; there are no hidden credential usages.
✓
Persistence & Privilege
always:false and no special persistence or system-wide modifications are requested. The skill does not attempt to modify other skills or global agent settings.
Assessment
This skill appears coherent for brand-voice work and contains only local Python scripts that analyze text and build system prompts. Before installing/use: (1) Only feed corpora you are allowed to share — scripts read any file path you provide. (2) Expect the generated system prompts to enforce silent replacement of 'prohibited' words and strong on-brand rewriting; always perform the documented Manual Review step before deploying prompts or publishing rewritten content. (3) Do not deploy generated system prompts to production without human oversight, since they can change meaning to keep 'voice'. (4) If you are concerned about privacy or sensitive data, inspect or sandbox the two Python scripts locally; they do not call external endpoints or use credentials.

Like a lobster shell, security has layers — review code before you run it.

latestvk9796g88exdcwvqbkqm5xbyk5d83fmx0

License

MIT-0
Free to use, modify, and redistribute. No attribution required.
Termshttps://spdx.org/licenses/MIT-0.html

Comments