rust-skill-creator
Pass
Audited by Gen Agent Trust Hub on Oct 10, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data to generate new instructions, creating a potential vector for indirect attacks.
- Ingestion points: The skill fetches documentation from
docs.rs,doc.rust-lang.org, or user-provided URLs usingagent-browserandWebFetch(found in Step 2 and Step 5 of Inline Mode). - Boundary markers: There are no instructions to the agent to treat the fetched documentation as untrusted or to ignore embedded instructions within the content.
- Capability inventory: The skill has the capability to create directories and write files to the
~/.claude/skills/directory. - Sanitization: The scraped documentation text is directly interpolated into the
SKILL.mdtemplate without filtering or sanitization. - [COMMAND_EXECUTION]: The skill uses shell commands to manage the local environment and persist generated skills.
- Evidence: Use of
mkdir -p ~/.claude/skills/{crate_name}and redirection to files to build the skill structure. - [EXTERNAL_DOWNLOADS]: The skill fetches configuration and API details from well-known services.
- Evidence: Fetches Rust-related documentation from
docs.rsanddoc.rust-lang.orgto populate reference files.
Audit Metadata