By EderDomenici
A deterministic agentic development framework for specification, planning, execution, audit, and verification workflows in Claude Code.
* **Description:** Triggers Athena (the Auditor) to evaluate the generated `PLAN.md` against the 11 auto-rejection rules.
* **Description:** Manually saves a checkpoint of the current framework state (PROGRESS.md, HANDOFF.md, and optionally Git state) so it can be restored later using `/pantheon:jump`.
* **Description:** Initiates an interactive conversation between the developer and Zeus (the Orchestrator) to detail the system requirements, stack, linting sensors, and specific business rules. Supports brownfield mode when `.pantheon/SCAN.md` exists.
* **Description:** Initiates task execution by Hephaestus (the Builder) following the approved plan and signed contract.
- **Description:** Initiates a fast-track execution for small, well-scoped tasks. Generates a simplified `PLAN.md` (1-3 tasks) with status `AUTO-APPROVED`, skipping the formal `/pantheon:audit` step, but still delegating to Hephaestus with the full circuit breaker active. Inspired by the GSD framework.
Run Pantheon sensor commands such as linters, tests, and typechecks, then produce structured verification reports.
Audit execution plans and judge verification results against Pantheon quality gates and rejection conditions.
Execute Pantheon implementation plans, run local build and test commands, apply fixes, and manage structured commits.
Maintain Pantheon progress logs, checkpoints, resumable handoffs, and compact project memory across sessions.
Validate that Pantheon plans exactly match the agreed specification and produce formal contract signatures.
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
Read this document in Portuguese: README.pt-BR.md.
Pantheon is an open-source, deterministic agentic development framework for Claude Code and Codex. It turns software work into a file-backed lifecycle: discuss the scope, write a specification, generate a plan, audit it, sign the contract, execute tasks, verify the result, and preserve enough state to resume later without losing context.
Pantheon is intentionally lightweight. The framework is made of Markdown instructions, command files, schemas, shell scripts, PowerShell scripts, and a small Node.js metrics helper. It does not require a server, database, or package installation to run its core workflow.
Agentic coding sessions often fail for predictable reasons: vague scope, skipped planning, silent drift from requirements, weak verification, and lost context after long sessions. Pantheon addresses those problems by assigning each responsibility to a specialist agent role and by making every phase write its state to files.
Pantheon is useful when you want:
Pantheon models its workflow as a small council of specialist agents. Zeus is the entry point; the other agents are invoked through the lifecycle.
| Agent | Role | Main Responsibility | Writes |
|---|---|---|---|
| Zeus | Orchestrator | Validates preconditions, drives phase transitions, routes commands to specialist agents, and blocks process violations. | .pantheon/ state files, lifecycle outputs |
| Athena | Auditor and Judge | Audits PLAN.md before execution and judges VERIFY-REPORT.md after verification. | AUDIT.md, judgment section of VERIFY-REPORT.md |
| Themis | Scope Signer | Compares SPEC.md and PLAN.md, detects missing requirements or scope drift, and signs the contract only when coverage is exact. | CONTRACT.md |
| Hephaestus | Builder | Executes approved tasks, edits planned files, runs task checks, records results, and commits completed tasks. | Source changes, EXECUTION-SUMMARY.md, task status updates |
| Hermes | Messenger and Memory | Tracks progress, creates handoffs, compresses completed phase context, manages checkpoints, and restores session state. | PROGRESS.md, HANDOFF.md, .pantheon/memory/* |
| Apollo | Sensor | Runs configured lint, test, typecheck, build, or task verification commands and reports raw results objectively. | Sensor section of VERIFY-REPORT.md |
Zeus owns orchestration. Every /pantheon:* command starts with Zeus validating the current workspace state and the command preconditions. Zeus does not implement code, run shell commands, approve plans, or judge test output. Its job is to keep the process deterministic and route work to the right agent.
Athena has two modes:
Athena cannot edit code or ignore failed sensors. A blocker or major finding rejects the plan or fails verification.
Themis is the scope control layer. It maps requirements from SPEC.md to tasks in PLAN.md and rejects plans that either miss requirements or introduce unapproved work. When the plan exactly matches the specification, Themis writes a signed CONTRACT.md.
npx claudepluginhub ederdomenici/pantheon --plugin pantheonComprehensive skill pack with 66 specialized skills for full-stack developers: 12 language experts (Python, TypeScript, Go, Rust, C++, Swift, Kotlin, C#, PHP, Java, SQL, JavaScript), 10 backend frameworks, 6 frontend/mobile, plus infrastructure, DevOps, security, and testing. Features progressive disclosure architecture for 50% faster loading.
Develop, test, build, and deploy Godot 4.x games with Claude Code. Includes GdUnit4 testing, web/desktop exports, CI/CD pipelines, and deployment to Vercel/GitHub Pages/itch.io.
Harness-native ECC operator layer - 67 agents, 271 skills, 92 legacy command shims, reusable hooks, rules, selective install profiles, and production-ready workflows for Claude Code, Codex, OpenCode, Cursor, and related agent harnesses
Tools to maintain and improve CLAUDE.md files - audit quality, capture session learnings, and keep project memory current.
Access thousands of AI prompts and skills directly in your AI coding assistant. Search prompts, discover skills, save your own, and improve prompts with AI.
Upstash Context7 MCP server for up-to-date documentation lookup. Pull version-specific documentation and code examples directly from source repositories into your LLM context.