convert-plugin-to-apm

Warn

Audited by Gen Agent Trust Hub on Jun 12, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes python scripts/validate_apm_package.py --path <target-path> via a shell. The use of <target-path> without explicit validation or sanitization instructions presents a risk of command injection if the path originates from an untrusted source or malicious project configuration.\n- [REMOTE_CODE_EXECUTION]: The skill references and executes a script at scripts/validate_apm_package.py. Because the content of this script is not included in the skill files provided, its behavior is unverifiable and could include dangerous operations like network exfiltration or unauthorized file system access.\n- [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection as it processes and migrates data from existing third-party plugins.\n
  • Ingestion points: Reads metadata from .claude-plugin/plugin.json and directory structures from skills/, agents/, and commands/.\n
  • Boundary markers: There are no markers or instructions provided to the agent to ignore potentially malicious instructions embedded in the source plugin's metadata or files.\n
  • Capability inventory: The skill has access to Bash (shell execution), Read (data ingestion), and Write (file creation) tools.\n
  • Sanitization: The skill does not define any sanitization or schema validation processes for the data it ingests before using it to generate new files or arguments.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Jun 12, 2026, 08:07 PM
Security Audit — agent-trust-hub — convert-plugin-to-apm