repo-scan
Fail
Audited by Gen Agent Trust Hub on Apr 1, 2026
Risk Level: HIGHEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The installation instructions direct the agent to fetch external code from an unverified GitHub repository: https://github.com/haibindev/repo-scan.git.
- [COMMAND_EXECUTION]: The skill uses a sequence of shell commands (git init, git fetch, git checkout, cp) to download and install this external code into the agent's local directory (~/.claude/skills/repo-scan).
- [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection as it is designed to scan and process untrusted external source code files.
- Ingestion points: Local source code files within repositories targeted for scanning.
- Boundary markers: Absent; there are no instructions to ignore or delimit embedded commands within the scanned files.
- Capability inventory: Extensive file system reading and generation of interactive HTML reports.
- Sanitization: Absent; the skill does not specify any validation or filtering of the content being analyzed.
Recommendations
- AI detected serious security threats
Audit Metadata