plugin-creator
Pass
Audited by Gen Agent Trust Hub on Oct 4, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill performs legitimate local project scaffolding operations. The provided scripts and instructions align with the stated purpose of creating and managing plugin manifests and directory structures. No suspicious network operations, credential access, or obfuscation techniques were found.
- [COMMAND_EXECUTION]: The skill documentation includes instructions for executing a local Python script (
scripts/create_basic_plugin.py). This script performs filesystem operations (creating directories and writing JSON files) necessary for its scaffolding functionality. The execution is restricted to local scripts provided within the skill package. - [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied input (plugin names and paths) that are written into local configuration files. The risk is mitigated by the following:
- Ingestion points: User-provided
<plugin-name>and path arguments in the shell commands described inSKILL.md. - Boundary markers: No explicit boundary markers are used in the prompt instructions to isolate data from instructions.
- Capability inventory: The skill has the capability to create directories and write to files using
pathlibandjson.dumpwithin thescripts/create_basic_plugin.pyscript. - Sanitization: The
normalize_plugin_namefunction uses a strict regular expression ([^a-z0-9]+) to filter the plugin name, converting it to lowercase hyphen-case and preventing injection of control characters or path traversal sequences into the filename.
Audit Metadata