kata-whats-new
Pass
Audited by Gen Agent Trust Hub on Feb 18, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
-
COMMAND_EXECUTION (SAFE): The skill uses the
catcommand to read theVERSIONandCHANGELOG.mdfiles from the$CLAUDE_PLUGIN_ROOTdirectory. This is standard behavior for a plugin to access its own metadata and is not considered a security risk. -
EXTERNAL_DOWNLOADS (LOW): The skill fetches content from
https://raw.githubusercontent.com/gannonh/kata/refs/heads/main/CHANGELOG.md. The GitHub usergannonhis not part of the trusted organizations list, making this an untrusted external source. -
PROMPT_INJECTION (LOW): (Category 8: Indirect Prompt Injection) The skill is vulnerable to instructions embedded in the remote changelog file.
- Ingestion points: The
fetch_remote_changelogstep usesWebFetchto download the remoteCHANGELOG.mdfile. - Boundary markers: Absent. The prompt provided to
WebFetch("Extract all version entries...") does not include delimiters or instructions to ignore embedded commands. - Capability inventory: The skill can perform local file reads (
cat) and network reads (WebFetch), and has the capability to provide formatted responses to the user. - Sanitization: Absent. There is no evidence of filtering or escaping the fetched content before it is processed by the agent.
- Ingestion points: The
Audit Metadata