Memory Management

Manage and standardize trading decision records, extract lessons, and support history retrieval and comparison within the PAI trading system.

This skill's purpose (trading-memory management) is plausible, but the SKILL.md expects two Python scripts (memory_evaluator.py and auto_memory_manager.py) and references an absolute home path (/Users/zst/clawd/...). Before installing or enabling this skill: 1) ask the author to provide the missing scripts or a clear, portable install plan; 2) ensure the skill uses relative or configurable paths rather than hard-coded absolute paths to avoid accidental access to your home directory; 3) review exactly which files it will read/write and whether you trust it to archive/delete files; 4) run it in a sandbox or on a test dataset first; and 5) decline or mark as untrusted if the author cannot explain or remove the hard-coded paths and supply the referenced tooling. Additional information that would change the assessment: included script files, an install spec that clearly installs vetted code, or removal of absolute user paths in favor of configurable locations.

Type: OpenClaw Skill Name: trading-memory-management Version: 1.0.0 The `SKILL.md` file instructs the AI agent to manage 'Related Files' using absolute paths (`/Users/zst/clawd/MEMORY.md`, `/Users/zst/clawd/memory/`). These paths point to a specific user's home directory, which is outside the typical sandboxed or relative working directory for a skill. This constitutes a prompt injection risk, as it directs the agent to interact with files in a sensitive user directory, potentially leading to unauthorized data access, modification, or deletion, even if the stated purpose is benign memory management.