html
Pass
Audited by Gen Agent Trust Hub on Aug 25, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
- [PROMPT_INJECTION]: The skill processes user-provided data to generate HTML artifacts without explicit sanitization or validation of the input content. This creates a surface for indirect prompt injection if the user provides malicious scripts or HTML tags.
- Ingestion points: User-provided HTML content specifications in SKILL.md.
- Boundary markers: Absent.
- Capability inventory: Writing to the local filesystem and starting a local web server (SKILL.md).
- Sanitization: Absent.
- [COMMAND_EXECUTION]: The skill executes a local web server using Python's built-in
http.servermodule. - Evidence:
python3 -m http.server "$PORT" --bind 127.0.0.1 --directory "$ARTIFACT_DIR"in SKILL.md. - [DATA_EXFILTRATION]: The skill writes generated artifacts to a directory within the user's home folder.
- Evidence:
Resolved ARTIFACT_DIR inside the provided HTML state root, or expanded ~/.html otherwisein SKILL.md.
Audit Metadata