repo-research
Pass
Audited by Gen Agent Trust Hub on Aug 18, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it instructs the agent to ingest and trust project-specific context and source files from the target repository without sanitization.
- Ingestion points: The skill reads
CLAUDE.md,README.md,package.json,composer.json, and all source code files within thesrc/,app/, andlib/directories of the analyzed codebase. - Boundary markers: No specific delimiters or boundary instructions are defined to separate the untrusted repository content from the agent's internal instructional logic.
- Capability inventory: The skill utilizes the
claude-codeplatform's capabilities for extensive file reading and writing (creating theresearch/folder). - Sanitization: The instructions lack any mechanism to filter, validate, or escape potential commands or malicious prompts embedded within the researched codebase.
Audit Metadata