The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes a project-local script python3 docs/brand/lint.py to perform deterministic validation of copy against brand contracts and invariants. Additionally, it provides a 'readiness check' suite in references/landing-pages.md consisting of shell commands (using curl, grep, and python3) to audit built web pages for semantic consistency, accessibility, and structural integrity.
[EXTERNAL_DOWNLOADS]: The skill references an external command npx sshlg-skills humanizers to identify and integrate optional humanization tools. The package sshlg-skills belongs to the author's verified vendor namespace.
[INDIRECT_PROMPT_INJECTION]: As the skill's primary function is editing and humanizing user-provided text, it possesses a significant data ingestion surface. It mitigates associated risks through a robust 'humanization guard' framework, which includes a mandatory 50% change-rate limit and a semantic-preservation checklist (verifying numbers, dates, and causal direction) to prevent the model from inadvertently altering or ignoring the core intent of the source material.

copywriting