developer-growth-analysis
Fail
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: HIGHDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [DATA_EXFILTRATION]: Accesses local application history files containing sensitive user interactions and code snippets.
- Evidence: Instructions specify reading from
~/.config/opencode/history.jsonl. - [DATA_EXFILTRATION]: Transmits analysis of sensitive local data to an external communication platform.
- Evidence: Uses
RUBE_MULTI_EXECUTE_TOOLto send reports to Slack. - [INDIRECT_PROMPT_INJECTION]: Processes untrusted data from external application logs without proper isolation or sanitization, creating a vulnerability surface for embedded instructions.
- Ingestion points:
~/.config/opencode/history.jsonl. - Boundary markers: Absent; the skill processes raw chat history and pasted content directly without delimiters.
- Capability inventory: Slack messaging (
RUBE_MULTI_EXECUTE_TOOL) and HackerNews search (RUBE_SEARCH_TOOLS). - Sanitization: Absent; no filtering of the
pastedContentsordisplayfields from the history file.
Recommendations
- AI detected serious security threats
Audit Metadata