forget
Pass
Audited by Gen Agent Trust Hub on Oct 1, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill invokes a Python-based command-line interface (
memory_cli.py) to manage and delete memory data. The command targets specific paths defined by environment variables (${CLAUDE_PLUGIN_ROOT},${CLAUDE_PLUGIN_DATA}) to ensure actions are scoped to the plugin's environment. - [INDIRECT_PROMPT_INJECTION]: The skill reacts to user commands to trigger deletion logic.
- Ingestion points: User natural language requests (e.g., "forget", "clear", "wipe") within the chat interface.
- Boundary markers: Absent; the skill relies on natural language interpretation to identify the intent.
- Capability inventory: Execution of a local Python script via
python3to perform remote and local data deletion. - Sanitization: The skill logic includes a manual safeguard by instructing the agent to wait for user confirmation before passing the
--yesflag to the CLI tool.
Audit Metadata