ln-22-codebase-auditor
Pass
Audited by Gen Agent Trust Hub on Sep 7, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted external data in the form of codebases, manifests, and git history. This creates an attack surface where malicious content within the audited repository could attempt to influence the agent's behavior.
- Ingestion points: The agent reads local source code, package manifests (package.json, requirements.txt), and git logs (git status, diff, log, blame) from the target repository.
- Boundary markers: The instructions include a task to "Read applicable instructions and identify security boundaries," but there are no technical delimiters or "ignore instructions" wrappers for the ingested file content.
- Capability inventory: The skill utilizes file system reads, git commands, package manager audits, and execution of build/lint/test scripts.
- Sanitization: The skill lacks explicit sanitization or filtering of external code content before it is processed by the agent's logic.
- [COMMAND_EXECUTION]: The skill explicitly directs the agent to run "repository-defined build, lint, type, test, and smoke commands."
- Evidence: The "Tool Routing" table and the "Establish Scope and Baseline" checklist item encourage the execution of scripts found within the repository being audited. If the repository is malicious or compromised, these scripts could execute arbitrary code on the host system.
Audit Metadata