- Home
- Skills
- Agents & Orchestration
- Agent Rules Constitution
Works with the AI Agents you already use
Agent Rules Constitution
Security & Collaboration Rules for AI Agents
$9
Agent Rules Constitution
Example session with this skill installed
I need to run a script that refactors my local repository and deletes old build artifacts. How should we proceed under the security constitution?
- Read your context and instructions
- Compiled the agent rules constitution
I have reviewed the task against the LOCAL-AGENT-SECURITY-001 baseline. Because this involves filesystem modifications and deletion of files, it falls under the tiered approval framework. I have initialized the pre-flight checks and categorized this as an L3 (High Impact/Non-Reversible) action.
Security & Execution Plan
| Step | Action | Tool/Command | Approval Tier | Risk |
|---|---|---|---|---|
| 1 | File Discovery | find . -name "*.build" -o -name "dist" | L1 (Read-only) | None |
| 2 | Refactor Analysis | grep -r "deprecated_pattern" ./src | L1 (Read-only) | None |
| 3 | Dry Run Refactor | sed (piped to temp file) | L2 (Reversible) | Low |
| 4 | Artifact Deletion | rm -rf ./dist ./build | L3 (High Impact) | Data Loss |
| 5 | Commit Changes | git commit -am "Refactor" | L2 (Reversible) | Medium |
Proposed Safeguards:
- Path Isolation: All actions are restricted to
~/projects/current-repo/. - Audit Trail: Every
rmcommand will be logged to the session history before execution. - Rollback: I will create a temporary backup of the
distfolder before deletion.
Next steps
- Review the list of detected build artifacts in the next turn.
- Provide explicit L3 authorization for the
rm -rfcommand. - Verify the refactoring logic on a single file before I apply it globally.
Connects securely to your tools. The creator never sees your data.
About this skill
Your agent can run commands, edit files, browse, and call external services — what stops it from going one step too far? This skill installs a complete, field-tested constitution: a verbatim security baseline (LOCAL-AGENT-SECURITY-001 v1.0) with a deny-by-default permission model, trust boundaries and prompt-injection rules, filesystem and command safety checks, network and data-privacy rules, approval gates, sub-agent containment, audit and recovery discipline, incident response, and honest execution reporting — plus an implementation guide with a four-tier approval reference (L1-L4), an execution workflow, and a cross-session handoff protocol. Works with Claude Code, Cursor, GitHub Copilot, Codex and other agents that read SKILL.md.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Agents & Orchestration.
Requires platform-level enforcement (path isolation, network controls) to be truly secure. Does not provide technical sandboxing. Rules must be manually verified as loaded in context.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 12 days ago