skill-edit

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill requires loading, converting, and preserving "100% of the data" from SKILL.md (including fetching upstream content) and printing before/after diffs in summaries, so any embedded secrets would need to be copied and emitted verbatim by the agent.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 1.00). Step 1 of SKILL.md explicitly instructs the agent to "search online broadly" and to fetch and diff upstream content (e.g., raw.githubusercontent.com/...) for github-imported skills, meaning the agent will ingest and act on public, user-generated web content that can influence edits and tool use.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill's runtime Step 1 explicitly instructs fetching upstream skill content from raw.githubusercontent.com/alirezarezvani/claude-skills/main/... and using that fetched content to diff/convert SKILL.md (i.e., it is fetched at runtime and can directly control the agent's instructions), so I flag that URL.