twg-context-discovery
Pass
Audited by Gen Agent Trust Hub on Aug 11, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute the
twg(Teamwork Graph) command-line tool. It specifies installation paths for the binary within user-writable directories such as$HOME/.local/bin/twgand%LOCALAPPDATA%\Programs\twg\bin\twg.exe. These are recognized as legitimate vendor resources associated with Atlassian.\n- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection due to its design, which involves aggregating and processing content from external, untrusted collaboration platforms.\n - Ingestion points: The agent ingests data from Jira work items, pull requests, project documents, search results, and meeting evidence as defined in
SKILL.md.\n - Boundary markers: The instructions lack explicit delimiters or specific markers to isolate ingested external data from the agent's core logic.\n
- Capability inventory: The skill uses the
twgCLI to perform deep graph queries and expand entity relationships based on the external data.\n - Sanitization: No explicit sanitization, filtering, or validation logic is provided for the content retrieved from external sources.
Audit Metadata