Manage versioned normative rules and operational specifications for AI agent governance.
npx skills add https://github.com/filipexyz/ravi --skill specs-filipexyz
Please help me install this Agent Skill. Skill: specs Source: https://github.com/filipexyz/ravi/tree/main/src/plugins/internal/ravi-system/skills/specs Command: npx skills add https://github.com/filipexyz/ravi --skill specs-filipexyz
This Skill solves the problem of agent drift and inconsistent behavior by providing a centralized, versioned memory for rules, runbooks, and invariants that agents must consult before executing tasks.
Use the specs skill to retrieve the current rules for the channels presence lifecycle feature.
High-intent search queries and answers about installing and using this skill.
Agent governance uses normative rules, operational specifications, and validation checks indexed in a versioned repository. This mechanism enforces system invariants and ensures agents act with full operational context before executing modifications.
To set up agent specs, establish a structured Markdown directory hierarchy as the source of truth and use SQLite for indexing. This allows you to retrieve specific operational modes like runbooks, checks, or decision history for governance.
Yes, SQLite is required to index the structured Markdown directory hierarchy. This setup enables efficient contextual retrieval of domain-specific constraints, runbooks, and validation checks that maintain the source of truth for system invariants.
Yes, contextual retrieval allows you to pull specific operational modes like runbooks, checks, and decision history. This ensures agents act with full context and prevents regressions when modifying core system components.
You need a centralized repository when agents must consult system invariants before executing tasks to prevent drift. It provides the versioned memory required for maintaining consistent decision-making across domains, capabilities, and features.
daymadeConsolidate redundant documentation files into a single source of truth.
maruakshayAudit CrewAI agent systems for role-boundary violations and prompt-injection risks.
caobingshengCoordinate dynamic AI context management across memory, tools, and retrieval systems.
ccfDesign, optimize, and govern production prompts with templates and verification steps.
