ste

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

• Third-party content exposure detected (medium risk: 0.30). The runtime workflow ingests free text only from the user’s chosen “source” content (e.g., merge-request/procedure/document text supplied in the conversation) and passes it directly to scripts/ste-check.sh for checking and annotation.

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

• Potentially malicious external URL detected (high risk: 0.80). The skill's Step 2 runs a runtime curl to download and parse https://www.asd-ste100.org/assets/files/ASD-STE100_ISSUE9.pdf into the checker cache (required for Verified/Marked operation), so remote content directly controls the checker's classification/outputs.