redmine-agent-ready-tickets
Pass
Audited by Gen Agent Trust Hub on Sep 6, 2026
Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted input from external Redmine issues which could contain malicious instructions designed to manipulate the agent.
- Ingestion points: The scripts retrieve full issue content including descriptions and comments in SKILL.md and redmine_ai_queue.py.
- Boundary markers: No specific delimiters or safety instructions are used to separate untrusted Redmine data from the agent's primary instructions.
- Capability inventory: The skill allows the agent to modify Redmine issues, create new tickets, and interact with the local file system.
- Sanitization: Incoming data from the Redmine API is processed without explicit sanitization or instruction filtering.
- [COMMAND_EXECUTION]: The agent is instructed to execute local helper scripts such as redmine_ai_queue.py and validate_ticket.py to perform its tasks.
- [DATA_EXFILTRATION]: The skill performs network operations to the user-specified REDMINE_URL to synchronize ticket data and implementation contracts.
Audit Metadata