Agent Policy — Development skill for Claude Code
One command policy for coding agents (Claude Code and Codex): severity-tiered allow/ask/deny rules, a shared guard hook, a sandbox layer and a scratchpad helper.
How to install Agent Policy
This entry records only its repository, not the path inside it, so there is no
exact command to give. Open ivan-pinatti-labs/agent-policy and copy the folder into
~/.claude/skills/, or the file into ~/.claude/agents/.
What Agent Policy does
One command policy for coding agents (Claude Code and Codex): severity-tiered allow/ask/deny rules, a shared guard hook, a sandbox layer and a scratchpad helper.
Alternatives in Development
- Obsidian Health — Run a vault health check — grouped by severity, detects contradictions, concept gaps, stale claims, and struct 277 ★
- Openshell Deepagent — A general-purpose coding agent that runs inside an NVIDIA OpenShell sandbox, orchestrated by Deep Agents and p 187 ★
- Incident Commander — Incident Commander - structured incident response with severity assessment and blameless post-mortems 127 ★
README
agent-policy
[](LICENSE.md) [](https://github.com/ivan-pinatti-labs/agent-policy/issues) [](https://github.com/sponsors/ivan-pinatti) [](https://github.com/ivan-pinatti-labs/agent-policy) [](https://github.com/ivan-pinatti-labs/agent-policy/forks) [](https://coderabbit.ai) [](https://sonarcloud.io/project/overview?id=ivan-pinatti-labs_agent-policy)
One command policy for coding agents, written once and installed for **Claude Code** (every profile) and **Codex**. Every command the policy lists carries a severity, and the severity decides what the agent does:
| Severity | Decision | Examples |
|---|---|---|
| read, low | allow | reading anyt |
Related Skills
Jev Brig
A Claude Code hook that judges Bash commands against a session policy level, and answers allow/ask/deny with a
Sentinel Hook
A pre-action gate for Claude Code agents: allow, ask or deny every shell command and file write, with each dec
Open Jev Approvals
Fail-closed binary (allow/deny) approval gate for Codex and Claude Code hooks — every intercepted tool call is
Claude Code Mini
Async Python implementation of the Claude Code agent harness — real Bash/Read/Glob/Grep tools, three-tier cach
Claude Perm Sim
Claude Code permission rule simulator: load the merged allow, ask and deny settings, show which rule decides a
Backstage
Backstage: a macOS menu-bar focus guard. Automation browsers and background agents stay backstage and ask befo
Related Agents
Permissions Manager
Manage tool permission rules. Use when user says 'allow', 'deny', or 'ask' for a command, wants to modify perm
Auth Policy Validator
Валидатор ZenStack access control policies. USE PROACTIVELY после изменений @@allow/@@deny для проверки безопа
Permission Analyst
Recommends allow/deny permission rules from denial telemetry and policy via PromptWise. Use to reduce permissi