vibe-code-cleanup
Warn
Audited by Gen Agent Trust Hub on Oct 26, 2026
Risk Level: MEDIUMCREDENTIALS_UNSAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [CREDENTIALS_UNSAFE]: The skill instructs the agent to read .env.local in SKILL.md, which typically contains sensitive secrets and credentials used for local development. Accessing these files exposes their contents to the AI's processing context.
- [INDIRECT_PROMPT_INJECTION]: The skill involves reading and processing various files within a project's codebase. Maliciously crafted content in these files could attempt to influence the agent's behavior during the cleanup process. 1. Ingestion points: Commands like grep, find, and cat are used to read codebase files as specified in SKILL.md. 2. Boundary markers: No explicit delimiters or instructions to ignore embedded commands are present when reading external data in SKILL.md. 3. Capability inventory: The agent has the ability to execute code via npm run build, npm test, and write changes to the filesystem during refactoring as instructed in SKILL.md. 4. Sanitization: There is no mention of sanitizing or validating the contents of the files in SKILL.md before they are processed by the agent.
- [COMMAND_EXECUTION]: The skill utilizes several shell commands and package execution tools such as npx, npm, grep, and find to perform its tasks in SKILL.md. While these are standard development tools, they provide a broad surface for command execution.
- [EXTERNAL_DOWNLOADS]: The use of npx for tools like ts-prune in SKILL.md involves downloading and executing packages from the npm registry at runtime without fixed versions, which could lead to the execution of untrusted code.
Audit Metadata