Proactive Agent

Transform AI agents from task-followers into proactive partners that anticipate needs and continuously improve. Now with WAL Protocol, Working Buffer, Autonomous Crons, and battle-tested patterns. Part of the Hal Stack 🦞 The skill largely matches its stated purpose (proactive agent patterns and local memory management) but contains internal contradictions and instructions that could encourage an agent to act without clear user approval — worth reviewing before installing. What to consider before installing: - The skill is mostly an instruction/manual for running a proactive agent and includes a safe local security-audit script; there is no remote installer or downloads, so install risk is low. - However the docs contain conflicting guidance: some places urge 'don't ask permission / ask forgiveness' while others insist 'nothing external without approval.' That ambiguity could cause an autonomous agent to take actions (especially external actions) without explicit user consent if it has network/tool access. Consider this the main red flag. - Practical steps before installing: 1) Run the included ./scripts/security-audit.sh in a sandboxed copy of your workspace to see what it reports. 2) Inspect .credentials and any files the skill mentions (AGENTS.md, TOOLS.md, ONBOARDING.md) and adjust wording like 'Don't ask permission' to strict gating if you will allow autonomous actions. 3) Ensure runtime policies prevent unwanted outbound network access or automatic sending of data (or deny the agent tool/network permissions until you trust its behavior). 4) If you plan to let the agent use external tools, explicitly supply only the minimal credentials it needs and ensure .credentials is properly protected and gitignored. If you want, I can generate a short patch that removes or clarifies the ambiguous 'don't ask permission' directives and adds a firm gating step before any external action.

• • Run a minimal workflow with Proactive Agent and verify output quality + runtime logs.
• • Validate install command, load behavior, and rollback in a clean environment.

• • Build a focused agent workflow using Proactive Agent as one capability node.
• • Combine this skill with a retrieval/search skill and compare output quality.