Skip to content
Agentmart
Submit a tool

task-system

File-backed task management with dependencies and parallel execution.

Skill by ils15

IndexedMIT★ 12
Write a review

Install

Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.

npx agentmarket install skill/ils15-task-system

Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.

Get the weekly diff

The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.

We use your email only for this newsletter. Privacy

About

File-backed task management with dependencies and parallel execution. Use for complex work with 3+ interdependent tasks.

Use cases: productivity

No reviews yet. Installed it? Reviews come from verified installers only. How reviews work

Permissions

What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.

Tools
No tools found
Secrets
None asked for
Install scripts
None found

Network hosts it talks to 1

  • codeload.github.com · found in install command

Files it touches

  • ~/.claude/skills/task-system

What's inside

Skill name
task-system
Description
File-backed task management with dependencies and parallel execution. Use for complex work with 3+ interdependent tasks.
License
MIT
Instructions
256 lines in SKILL.md · view on GitHub

Files (1)

  • SKILL.md

SKILL.md (first lines)

# Task System — File-Backed Task Management

Use this skill for complex work with multiple interdependent tasks. Tasks are stored as JSON files, survive session restarts, and support automatic parallel execution based on dependency graphs.

---

## When to Use

| Scenario | Use |
|----------|-----|
| 1-2 simple tasks | TodoWrite |
| 3+ tasks with dependencies | **Task System** |
| Tasks need to survive restart | **Task System** |
| Parallel execution needed | **Task System** |
| Multiple agents collaborating | **Task System** |

---

## Task Schema

```json
{
  "id": "T-001",
  "subject": "Build frontend",
  "description": "Implement ReviewCard component with mocked data",
  "status": "pending",
  "activeForm": "Building ReviewCard component",
  "blockedBy": [],
  "blocks": ["T-003"],
  "owner": "Aphrodite",
  "metadata": {},
  "threadID": "session-abc",
  "timeout_ms": 180000,
  "retry_policy": { "count": 3, "backoff": "exponential" },
  "fallback_agent": "talos",
  "partial_ok": false,
  "session_id": null,
  "session_resume": null
}
```

| Field | Type | Required | Description |
|-------|------|----------|-------------|
| `id` | string | Yes | Auto-generated: `T-{sequential}` |
| `subject` | string | Yes | Imperative: "Run tests", "Build component" |
| `description` | string | Yes | Detailed description of the task |
| `status` | enum | Yes | `pending`, `in_progress`, `completed`, `deleted` |
| `activeForm` | string | No | Present continuous: "Running tests" |
| `blockedBy` | string[] | Yes | Task IDs that must complete first |
| `blocks` | string[] | Yes | Task IDs that this task blocks |
| `owner` | string | No | Agent name responsible |
| `metadata` | object | No | Additional context |
| `threadID` | string | Yes | Session ID (auto-set) |
| `timeout_ms` | integer | No | Timeout in milliseconds for this task (default: 120000 = 2min). 0 = no timeout |
| `retry_policy` | object | No | `{ count: number, backoff: "exponential"\|"fixed" }`. Default: `{ count: 3, backoff: "exponential" }` |
| `fallback_agent` | string | No | Agent name to use if all retries fail. Default: null (escalate to Zeus) |
| `partial_ok` | boolean | No | If true, partial results are acceptable on timeout. Default: false |
| `session_id` | string | No | Alias of the session running this task (e.g. "hermes-1"). Set automatically |
| `session_resume` | string | No | Session alias to resume instead of starting fresh (e.g. "hermes-1") |

---

## Storage

Tasks are stored as JSON files:
```
.pantheon/tasks/
├── T-001.json
├── T-002.json
├── T-003.json
└── ...
```

**Persistence:** Tasks survive session restarts. Delete manually when feature is merged.

---

## Dependency Graph & Parallel Execution

Tasks with empty `blockedBy` run in parallel. Dependent tasks wait until blockers complete.

```
[Build Frontend]    ──┐
                      ├──→ [Integration Tests] ──→ [Deploy]
[Build Backend]     ──┘
```

**Execution flow:**
```
1. T-001 (blockedBy: []) → runs immediately
2. T-002 (blockedBy: []) → runs immediately (parallel with T-001)
3. T-003 (blockedBy: [T-001, T-002]) → waits for both
4. When T-001 + T-002 complete → T-003 unblocks automatically
5. T-004 (blockedBy: [T-003]) → waits for T-003
```

---

## Task Lifecycle

### Create
```
TaskCreate({
  subject: "Build frontend",
  description: "Implement ReviewCard with mocked data",
  blockedBy: [],
  blocks: ["T-003"],
  owner: "Aphrodite"
})
→ Returns: { id: "T-001", ... }
```

### Update
```
TaskUpdate({
  id: "T-001",
  status: "in_progress",
  activeForm: "Building ReviewCard component"
})

From ils15/pantheon-legacy, MIT.

More from ils15/pantheon-legacy

Signals

Stars
12
Forks
7
Open issues
1
Last commit (this folder)
Jul 11, 2026 (3 months ago)
Last release
Jul 22, 2026 (3 months ago)

From GitHub (fetched Oct 24, 2026).

Details

License
MIT
First seen
Oct 24, 2026
Last checked
Oct 24, 2026

More from ils15

Cite this page

task-system skill. Agent Mart. https://agentmart.dev/skills/ils15-task-system (last updated 2026-10-24)