codebase-recon
Pass
Audited by Gen Agent Trust Hub on Sep 13, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses multiple git-based shell commands to extract metadata from the local repository, including commit logs, file change counts, and contributor statistics. These commands are executed locally to generate a project health report.\n- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from the repository's git history, such as commit messages and filenames, which are then processed and summarized by the agent. This represents an attack surface where an attacker could commit specifically crafted strings into the repository history to influence the agent's analysis or the content of the generated report.\n
- Ingestion points: Output from git log, git shortlog, and git branch commands in Phase 2.\n
- Boundary markers: None; the skill does not instruct the agent to use delimiters or ignore embedded instructions when reading git metadata.\n
- Capability inventory: The skill utilizes shell command execution (git subprocesses) and has the capability to write the final report to the local filesystem (e.g., docs/codebase-recon-report.md).\n
- Sanitization: The skill lacks explicit sanitization, validation, or escaping of the ingested repository metadata before it is interpolated into the final report summary.
Audit Metadata