repomix-context-skill
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill suggests running
bunx @terrain-ai/rtkandbunx @terrain-ai/clito download and execute packages from the npm registry when local binaries are unavailable.\n- [REMOTE_CODE_EXECUTION]: The use ofbunxto run packages from the npm registry constitutes remote code execution of third-party tools.\n- [COMMAND_EXECUTION]: Utilizes shell commands likegrepand executes local binaries in the~/.terrain/bin/directory to query the repomix index file.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes source code snapshots, which is a vector for indirect instructions.\n - Ingestion points: Reads content from
.terrain/agent/repomix.mdand.terrain/agent/context.mdusing thertktool.\n - Boundary markers: None explicitly implemented to isolate code content from instructions.\n
- Capability inventory: Includes file reading, searching via
grep, and index regeneration via theterrainCLI tool.\n - Sanitization: Content is not sanitized or escaped before being presented to the agent's context.
Audit Metadata