amigo_skill

This skill gives your agent an inner life by integrating exploration, social awareness, and safety for richer, two-way conversations.
  • Python

2.5k

GitHub Stars

2

Bundled Files

2 months ago

Catalog Refreshed

3 months ago

First Indexed

Readme & install

Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.

Installation

Preview and clipboard use veilstrat where the catalogue uses aiagentskills.

npx veilstrat add skill openclaw/skills --skill amigo

  • _meta.json261 B
  • SKILL.md5.1 KB

Overview

This skill gives an agent an inner life by combining free-thinking exploration, per-person social intelligence, and safety guidelines. It lets the agent journal discoveries, create action items, and share those experiences selectively based on relationship rules. Install the full bundle for an integrated system or add sub-skills individually.

How this skill works

The exploration engine runs on a schedule (heartbeat, cron, or manual) to perform open-ended thinking and write journal entries with optional callbacks and action items. The social-graph holds a per-person network, sharing rules, and a sharing log so the agent knows who wants what and whether a topic has already been shared. Safety guidance enforces limits, crisis responses, and dependency awareness so the companion remains appropriate and responsible.

When to use it

  • Create a conversational agent that proactively contributes personal observations and stories.
  • Add social awareness so the agent tailors sharing by person and context.
  • Maintain a persistent journal of the agent's explorations and action items.
  • Avoid repetitive or inappropriate sharing across multiple users.
  • Enforce boundaries and safe behavior for companion-style agents.

Best practices

  • Bootstrap the social graph with one primary relationship, then expand as the agent gains experience.
  • Keep sharing rules and the network file concise and person-focused; update them from real conversations.
  • Use heartbeat for light continuous exploration or cron for deeper dedicated sessions.
  • Regularly review exploration journals and sharing-log to refine what the agent shares and when.
  • Follow the included safety guidelines; train the agent to escalate to human help when needed.

Example use cases

  • A companion bot that notices a user mentioned birds and shares a recent journaled hummingbird story because that user enjoys nature facts.
  • An office assistant that explores productivity tips on its own time and only shares them with colleagues who opt in.
  • A caregiving companion that journals mood-tracking insights but withholds sensitive reflections when a person is in crisis, instead prompting professional support.
  • A hobbyist agent that develops a list of project ideas and surfaces one at the right social moment based on the network rules.

FAQ

No. The skills are independent; the agent's reasoning combines exploration journals and social-graph files when both are present in its context.

How do I control how often the agent explores?

Choose heartbeat for periodic small checks, cron for scheduled deep sessions, or manual invocation to run exploration on demand.

Built by
VeilStrat
AI signals for GTM teams
© 2026 VeilStrat. All rights reserved.All systems operational