Second Order Effects

Analyze decisions by tracing consequences beyond immediate outcomes to second and third-order effects.

This skill is internally coherent and appears safe from a capability-misalignment perspective, but review these practical points before installing: - Local storage: it will create and write files under ~/second-order-effects/ (memory.md, decisions/, patterns.md). Those files may contain sensitive decision details — inspect and secure that directory if you care about privacy. - Persistence: decision logs are append-only by design; if you want deletions or expiration, plan a manual cleanup or change the workflow. - External data/network: the SKILL.md promises not to send data externally, but that is a behavioral claim in prose. If you need a stronger guarantee, confirm that the agent runtime enforces network/file access restrictions for skills before trusting highly sensitive content. - Permissions: consider setting restrictive filesystem permissions on ~/second-order-effects/ if multiple users share the machine. If these trade-offs are acceptable (local journaling of analyses), the skill is appropriate to install. If you prefer no local persistence or need provable network isolation, either modify the skill or avoid installing it.

Type: OpenClaw Skill Name: second-order-effects Version: 1.0.0 The OpenClaw skill bundle 'second-order-effects' is classified as benign. All files (SKILL.md, framework.md, memory-template.md, patterns.md, setup.md) consistently describe a local analytical tool for an AI agent, focusing on decision analysis and local data storage. SKILL.md explicitly states that the skill 'does NOT: Send data externally, Access files outside its directory, Make network requests', and the `_meta.json` confirms no external binaries are required. There is no evidence of prompt injection attempts, malicious execution, data exfiltration, or any other harmful behavior across the analyzed content.