OmniTranslate

Apply high-fidelity localization to structured artifacts such as web apps, docs, PDFs, slide decks, Office files, subtitles, code repositories, and game asse...

Security Scan

The skill's files and runtime instructions match its stated purpose (high-fidelity localization); the included probe script and documentation are coherent and there are no disproportionate permissions, installs, or hidden endpoints.

Purpose & Capability

Name/description align with the provided assets: a lightweight probing script plus detailed format/pipeline docs used to plan structured localization. There are no unrelated required env vars, binaries, or config paths.

Instruction Scope

SKILL.md instructs the agent to run scripts/probe_artifacts.py on a user-specified path and follow the included reference docs. The instructions do not reference extraneous files, network endpoints, or credentials. The probe script only enumerates and classifies files under the given path (no network I/O).

Install Mechanism

No install specification is present and there are no downloads or package installs. This is instruction-only with one small Python utility, which minimizes on-disk and supply-chain risk.

Credentials

The skill requests no environment variables or credentials (proportional). Note: the probe script reads the filesystem at whatever path the user/agent supplies — if run against a sensitive directory it can enumerate filenames and categories. That behavior is consistent with its purpose but worth being careful about what paths are supplied.

Persistence & Privilege

The skill does not request persistent or elevated privileges, does not set always:true, and does not modify other skills or global agent config. Autonomous invocation is allowed by platform default but is not combined with broad credentials or other red flags here.

Assessment

This skill appears internally consistent and the included probe script is readable and purpose-built. Before installing or running it: 1) review scripts/probe_artifacts.py yourself (it only walks the filesystem and classifies files); 2) run the probe on a non-sensitive sample folder first (avoid pointing it at root or directories with secrets); 3) confirm you trust the skill source/owner since there is no homepage; and 4) if you are concerned about an agent invoking it autonomously, restrict use or disable autonomous invocation in agent settings. If you want extra assurance, run the script in a sandboxed environment or inspect the repository for any changes before use.

Quick Start

1. Run python scripts/probe_artifacts.py <path> on the target file or folder.
2. Read references/workflow.md.
3. Apply references/decision-thresholds.md before editing.
4. Use references/artifact-pipelines.md to select the correct path.
5. Use references/translation-boundaries.md to protect non-translatable content.
6. Load references/format-risk-checklists.md for the active format.
7. Use references/locale-sensitive-typography.md for CJK, RTL, or mixed-script output.
8. Validate with references/quality-gates.md before handoff.

Operating Rules

• Prefer editable sources over derived exports.
• Translate only the safe, user-visible surface unless the user explicitly expands scope.
• Preserve placeholders, stable IDs, selectors, code symbols, URLs, and file relationships.
• Fail closed when safe round-tripping is not possible.
• Report skipped content and residual risk explicitly.

Deliverables

• the localized artifact or repository
• a concise note describing the selected pipeline
• a list of protected or skipped items
• a validation summary with any remaining fidelity risk