Oracle
- 7 repo stars
- Updated June 21, 2026
- andreasasprou/agent-skills
Oracle is an agent skill that provides read-only GPT-5.2 strategic reasoning for architecture, security analysis, and complex debugging.
About
Oracle is an agent skill that plugs in a GPT-5.2 strategic advisor with emphasis on deep reasoning. It is built for developers who need a careful second brain on architecture decisions, security implications, and gnarly debugging—without granting write access that could accidentally change the codebase. Use it when you are stuck between design options, unsure about a security boundary, or need structured hypotheses before you instrument or patch production issues. Because the skill is read-only by design, it complements your primary coding agent: you stay in the driver's seat while Oracle pressure-tests assumptions and surfaces risks. It is not a substitute for automated scanners or penetration tests, but it shines as a reasoning layer across build-time design, pre-ship security review, and incident analysis when you want narrative clarity before you act.
- GPT-5.2 deep-reasoning advisor for architecture, security analysis, and complex debugging
- Read-only posture so recommendations do not auto-edit files or run destructive commands
- Suited to high-stakes design forks, threat modeling, and root-cause hypotheses
- Fits solo builders who want a senior-style review pass before committing to a path
Oracle by the numbers
- Data as of Jul 7, 2026 (Skillselion catalog sync)
/plugin marketplace add andreasasprou/agent-skills/plugin install oracle@andreas-agent-skillsAdd your badge
Show developers this skill is listed on Skillselion. Paste this into your README.
| repo stars | ★ 7 |
|---|---|
| Last updated | June 21, 2026 |
| Repository | andreasasprou/agent-skills ↗ |
What it does
Get a read-only GPT-5.2 strategic second opinion on architecture, security tradeoffs, and hard debugging without letting the advisor mutate your repo.
Who is it for?
Best when you're debating stack or module boundaries, reviewing security before a release, or untangling multi-layer bugs with a dedicated reasoning pass.
Skip if: Skip if you need hands-on refactors, automated remediation, or verified audit attestations without human follow-through.
When should I use this skill?
You need strategic GPT-5.2 reasoning on architecture, security, or debugging with read-only agent access.
What you get
You leave with clearer tradeoffs, threat angles, and debugging directions you can implement with your main agent in controlled, explicit steps.
- Structured recommendations and tradeoff analysis
- Security or debugging hypotheses for you to implement manually
Files
GPT-5.2 strategic advisor with deep reasoning. Read-only access for architecture decisions, security analysis, and complex debugging.
oracle
Source: ./plugins/oracle
{ "name": "oracle", "source": "./plugins/oracle", "description": "GPT-5.2 strategic advisor with deep reasoning. Read-only access for architecture decisions, security analysis, and complex debugging." }
Related skills
Setup Matt Pocock SkillsScaffold the per-repo configuration that Matt Pocock’s engineering agent skills rely on so they understand the issue tracker, triage labels, and domain documentation la495k196k
Lark AppsConnect Claude, Cursor or custom agents directly to Lark (Feishu) for messaging, document automation, approval workflows and enterprise data access.402k
CavemanSlash token usage by roughly 75% while keeping every technical detail intact when working with Claude Code, Cursor or similar agents.391k94.7k
Lark Skill MakerQuickly turn any Lark/Feishu OpenAPI call or multi-step workflow into a reusable agent skill with its own SKILL.md.386k16k
Running Claude Code Via Litellm CopilotRun Claude Code at a fraction of the cost by routing requests through LiteLLM to the GitHub Copilot Chat API.270k72
Codex PetGenerate a complete Codex Pet spritesheet and metadata from one reference image without needing an OpenAI key or Codex Pro.246k10How it compares
Use as a read-only reasoning advisor instead of handing the same task to a write-capable agent with no guardrails.
FAQ
Who is oracle for?
Developers using Claude Code, Cursor, or Codex who want GPT-5.2-level strategic advice on design, security, and debugging without automatic code changes.
When should I use oracle?
During Build when choosing architecture or agent workflows; in Ship when reasoning about security boundaries before launch; and in Operate when diagnosing complex failures—whenever you need depth before you edit.
Is oracle safe to install?
Read-only access is the stated design, but you should still review the Security Audits panel on this Prism page and confirm plugin permissions in your agent before enabling it in sensitive repos.