Skip to content
Back to search

agentic-os

Agentic personal OS to automate high-leverage workflows with Codex, Claude Code, Pi, OpenClaw and other coding agents/ runtime platforms.

Safety scan

A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.

Checked today
Badge for your README
findsafeskills safety rating

Markdown

[![findsafeskills safety rating](/api/listings/itseffi--agentic-os/badge.svg)](/listings/itseffi--agentic-os)

HTML

<a href="/listings/itseffi--agentic-os"><img src="/api/listings/itseffi--agentic-os/badge.svg" alt="findsafeskills safety rating" /></a>

Install

i

Install instructions from the README

As written by the repo's author. findsafeskills has not verified these steps.

1. **Clone this repo**

   git clone https://github.com/itseffi/personal-os.git
   cd personal-os

2. **Run setup**

   chmod +x setup.sh
   ./setup.sh

3. **Start using** This automates high-leverage execution end-to-end: it converts raw backlog into prioritized, goal-aligned, verification-enforced action plans.

   Open this repo in your agent and run:
   1) "Process my backlog from BACKLOG.md into Tasks/**/*.md using AGENTS.md rules."
   2) "Show my P0/P1 unblocked tasks aligned to GOALS.md."
   3) "Propose today’s top 3 with required verification evidence and commands."

---

What was scanned

The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.

README.md · README · 197 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

  1. ![Agentic Personal OS Banner](Resources/assets/hero-banner-agentic-os.png)
  2. **TL;DR**: An agentic personal operating system built to automate high-leverage workflows across Claude Code, Codex, Pi, OpenClaw, and other coding agents/runtime platforms.
  3. ---
  4. ## Quick Start
  5. 1. **Clone this repo**
  6. ```bash
  7. git clone https://github.com/itseffi/personal-os.git
  8. cd personal-os
  9. ```
  10. 2. **Run setup**
  11. ```bash
  12. chmod +x setup.sh
  13. ./setup.sh
  14. ```
  15. 3. **Start using**
  16. This automates high-leverage execution end-to-end: it converts raw backlog into prioritized, goal-aligned, verification-enforced action plans.
  17. ```
  18. Open this repo in your agent and run:
  19. 1) "Process my backlog from BACKLOG.md into Tasks/**/*.md using AGENTS.md rules."
  20. 2) "Show my P0/P1 unblocked tasks aligned to GOALS.md."
  21. 3) "Propose today’s top 3 with required verification evidence and commands."
  22. ```
  23. ---
  24. ## Quick Links
  25. [Build Your Personal OS](Tutorials/build-your-personal-os.md) · [Workflows](Workflows/README.md) · [Canonical Skills](.agents/skills/README.md) · [Evals](Evals/README.md) · [Tutorials (index)](Tutorials/README.md)
  26. ---
  27. ## Architecture
  28. ```mermaid
  29. flowchart TD
  30. U["User Prompt"] --> A["Agent Runtime<br/>Claude Code | Codex | Pi | OpenClaw"]
  31. A --> I["Instructions<br/>AGENTS.md + wrappers"]
  32. A --> S["Skills<br/>.agents/skills/*/SKILL.md"]
  33. A --> W["Workflows<br/>Workflows/*.md"]
  34. A --> F["State + Context<br/>Tasks, GOALS, BACKLOG, Knowledge, Resources"]
  35. A -. optional .-> M["MCP Integrations<br/>System/mcp + external services"]
  36. A -. optional .-> D["Subagents<br/>runtime-dependent delegation"]
  37. A --> E["Evals<br/>Evals/ + Evals/skills + scripts/run_skill_evals.py"]
  38. classDef core fill:#ff9891,stroke:#2b2b2b,color:#111111,stroke-width:1.2px;
  39. classDef optional fill:#ffd4d0,stroke:#2b2b2b,color:#111111,stroke-width:1.2px,stroke-dasharray: 4 3;
  40. class U,A,I,S,W,F,E core;
  41. class M,D optional;
  42. ```
  43. ---
  44. ## Agent Compatibility
  45. Personal OS is designed to work with Claude Code, Codex, Pi, OpenClaw, and similar coding agent runtimes.
  46. - Shared behavior: `AGENTS.md`
  47. - Claude wrapper: `CLAUDE.md`
  48. - Codex wrapper: `CODEX.md`
  49. - Pi wrapper: `PI.md`
  50. - OpenClaw wrapper: `OPENCLAW.md`
  51. - Canonical runtime skills: `.agents/skills/*/SKILL.md`
  52. - Skills in this repo follow the [Agent Skills open standard](https://agentskills.io/home).
  53. - This repo uses skills with progressive disclosure to manage context efficiently: agents begin with each skill's metadata (`name`, `description`, file path, plus `agents/openai.yaml`), and load full `SKILL.md` instructions only when a skill is selected. Canonical skills live in `.agents/skills/`, with bridge paths for Claude, Pi, and OpenClaw.
  54. - Optional subagents are supported when the runtime provides agent delegation features (not required for core repo operation).
  55. - Claude bridge path: `.claude/skills -> ../.agents/skills` (symlink)
  56. - Pi bridge: configure Pi to point to this repo and use `.agents/skills/` as its skill source
  57. - OpenClaw bridge: create `skills -> .agents/skills` symlink (or load `.agents/skills` via OpenClaw config)
  58. Bridge bootstrap (run once from repo root):
  59. ```bash
  60. mkdir -p .claude
  61. ln -sfn ../.agents/skills .claude/skills
  62. ln -sfn .agents/skills skills
  63. ```
  64. For Codex/OpenAI-style routing metadata, this repo includes:
  65. - `.agents/skills/<skill>/agents/openai.yaml`
  66. (Claude, Pi, and OpenClaw primarily use `SKILL.md` and do not require this file format.)
  67. ---
  68. ## Pi Local/Offline Setup (Optional)
  69. You can run Personal OS with Pi using a local/offline model backend (for example `llama.cpp`) or a hosted endpoint. For full setup instructions (server launch, `~/.pi/agent/models.json`, and runtime configuration), see [Pi Agent Setup](Tutorials/pi-agent-setup.md).
  70. ---
  71. ## File System Layout
  72. ```
  73. personal-os/
  74. ├── AGENTS.md # AI agent instructions (the brain)
  75. ├── GOALS.md # Your goals and priorities
  76. ├── BACKLOG.md # Quick capture inbox
  77. ├── Tasks/ # Your active work
  78. ├── Knowledge/ # Your notes and docs
  79. ├── Resources/ # Voice samples, templates, references
  80. ├── Workflows/ # Daily + Product & Strategy workflows
  81. ├── .agents/skills/ # Canonical Codex/OpenAI skill packs
  82. ├── Evals/ # Session reviews
  83. ├── Tutorials/ # Learning guides
  84. └── System/ # MCP server, templates, integrations
  85. ```
  86. Semantics by location: `Tasks/**/*.md` = actionable work, `Knowledge/**/*.md` = reference context.
  87. ---
  88. ## How It Works
  89. ### The Memory Stack
  90. ```
  91. AGENTS.md → Instructions layer (how AI behaves)
  92. GOALS.md → Priority layer (what matters)
  93. Tasks/**/*.md → State layer (current work)
  94. Knowledge/**/*.md → Context layer (reference)
  95. .agents/skills/* → Capability layer (how the agent executes specialized workflows)
  96. ```
  97. ### Privacy First
  98. Personal operating data stays local (gitignored):
  99. - `Tasks/` - your work
  100. - `Knowledge/` - your notes
  101. - `Resources/` - your samples
  102. - `BACKLOG.md` - your inbox
  103. Some top-level configuration files (`AGENTS.md`, `GOALS.md`, `CLAUDE.md`, `CODEX.md`, `PI.md`, `OPENCLAW.md`, docs) are version controlled by design. Treat `GOALS.md` as potentially sensitive and review content before publishing a public repo.
  104. ---
  105. ## Evals
  106. This repo includes structural, behavioral, routing, and memory-impact evals.
  107. Run:
  108. ```bash
  109. python scripts/validate_skills.py
  110. python scripts/validate_skill_eval_cases.py
  111. python scripts/run_skill_evals.py --provider fixture
  112. python scripts/run_routing_evals.py
  113. python scripts/run_memory_impact_evals.py
  114. ```
  115. Optional live-model run (OpenAI-compatible endpoint, local or remote):
  116. ```bash
  117. python scripts/run_skill_evals.py --provider openai --model your-model-id
  118. ```
  119. Outputs are written to:
  120. - `Evals/skills/results/`
  121. - `Evals/memory/results/`
  122. Use these evals as a regression gate when updating `.agents/skills/`.
  123. ---
  124. ## Long-Running Agent Principles
  125. Personal OS follows four operating patterns:
  126. - **Skills**: versioned procedures in `.agents/skills/*/SKILL.md`
  127. - **Shell execution**: run real tasks in terminal environments and produce artifacts
  128. - **Compaction-aware workflows**: structure long runs to preserve continuity
  129. - **Verification-first completion**: require fresh evidence before claiming work is done
  130. Security defaults:
  131. - Keep network access minimal and allowlist-based
  132. - Treat tool output as untrusted input
  133. - Use explicit review boundaries for generated artifacts
  134. ---
  135. ## Tech Stack
  136. - **File Format:** Markdown with YAML frontmatter
  137. - **Agent Runtimes:** Claude Code, Codex, Pi, OpenClaw, Cursor, and similar coding agent runtimes
  138. - **Optional Integrations:** MCP (Slack, Linear, Google Calendar, Atlassian, Granola)
  139. - **Version Control:** Git
  140. ---
  141. ## Contributing
  142. Issues and PRs welcome.

Ratings and comments

Sign in to rate this skill.

Sign in

Free to use. If it’s helped, pass it on.