Security audit
Security checks for vulnerabilities and agentic risk
This skill is a scoped microtypography helper that edits text formatting and only writes files or Figma nodes after user permission.
Before installing, be aware that the skill can change text by inserting invisible or subtle Unicode spacing characters and, when given Figma access, can edit selected text nodes after you approve it. Use the change log to review those edits, especially in multilingual content or systems with strict copy/localization rules.
The 'When To Use' section includes the trigger phrase 'polish this text', which is a generic editing request rather than a narrowly scoped microtypography invocation. Although other examples are specific, this broad phrase could cause the skill to activate for ordinary proofreading or copyediting requests outside its intended scope.
Skill enumerates or reads other installed skills. Access to other skills' SKILL.md files or the skills directory reveals prompt instructions, capabilities, and secrets that should be invisible to peer skills.
# Microtypography skill — test run
Manual dry-run of `skills/microtypography/SKILL.md` against four cases: PL reflowing markdown, EN reflowing markdown (strict mode), Case A hard-wrapped rewrap, Case B soft-hyphen + CSS suggestion. Run by following the skill's own Steps 1–5 and Quality Checklist.
**Notation used below** (source markdown would contain the real invisible characters, not these glyphs):
- `␣` = non-breaking space (U+00A0)
The input specification says to 'detect per-paragraph if mixed,' which imposes language handling automatically rather than offering the user a language/locale choice. This is a natural-language locale policy concern because the skill decides locale behavior on the user's behalf unless ambiguity is detected later.
Step 1 instructs the skill to 'Detect language per paragraph (or ask if ambiguous),' which makes language selection automatic by default. This can conflict with organizational language/locale choice expectations because the user is not given an explicit choice unless the system finds ambiguity.