Skills DirectorySkills Directory
SkillsLearnSecurityCategoriesDocsCommunityBlog
Sign InSubmit Skill
Skills Directory

Security-tested agent skills for Claude, coding agents, and AI workflows.

Directory

  • Browse Skills
  • All Skills A–Z
  • Claude Skills
  • Claude Code Skills
  • Agent Skills
  • Categories
  • Authors
  • Submit a Skill

Learn

  • Learn Hub
  • Install Claude Skills
  • Write SKILL.md
  • Skills vs MCP
  • Directories Compared

Security

  • Security
  • Methodology
  • Secure Claude Skills
  • Security Badges

Company

  • About
  • Community
  • Blog
  • API Docs
  • Advertise

2026 Skills Directory. All rights reserved.

ProTermsPrivacyRefunds
Back to skills

Orch Tasks

ASecurity

Use when specs under specs/ are ready to become tasks in an orch project (step 4 of orch-plan, or on its own after editing a spec). Runs orch atomize as a dry-run, shows the diff, applies it only after the user confirms, then makes the DAG valid with orch router and orch validate and explains how to promote tasks out of backlog.

2 stars
0 votes
0 copies
0 views
Added 9/23/2026
ai-agentsgobashbackend

Works with

cli

Security Analysis

A100/100

Scanned 9/23/2026

Install to Claude Code

$npx -y skills add hectorcanaimero/orch --skill orch-tasks --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Orch Tasks?

Add the live security badge to your README — it updates automatically with every re-scan.

Security grade badge for Orch Tasks
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/hectorcanaimero-orch-tasks/badge)](https://www.skillsdirectory.com/skills/hectorcanaimero-orch-tasks)

More formats (shields.io, HTML) on the badges page.

Download with Pro
Files
SKILL.md
---
name: orch-tasks
description: Use when specs under specs/ are ready to become tasks in an orch project (step 4 of orch-plan, or on its own after editing a spec). Runs orch atomize as a dry-run, shows the diff, applies it only after the user confirms, then makes the DAG valid with orch router and orch validate and explains how to promote tasks out of backlog.
---

# orch-tasks — spec → tasks.json

Last stage of **PRD → ARCH → SPEC → TASKS**, and the only one that changes
what orch will run. Everything here goes through the `orch` binary — never
edit `tasks.json` by hand, and never apply a change the user has not seen.

## 1. Preview

```bash
orch atomize                            # every spec under the spec root
orch atomize --file specs/f1-<slug>.md  # or just one
```

Without `--apply` this writes nothing. It prints the merge diff, with its
headings in Spanish:

- `Parser warnings` — lines the parser dropped or misread;
- `+ NUEVAS` — ids not yet in `tasks.json`;
- `~ ACTUALIZADAS` — title, description, model, reason, dependencies,
  estimate, specRef or phase changed in the spec, one line per field;
- `= SIN CAMBIOS` — a count, nothing to review;
- `⚠ HUÉRFANAS en tasks.json` — tasks no scanned spec mentions. Atomize
  **never deletes** them. With `--file`, every task from other specs shows up
  here simply because it was not scanned; that is expected.
- `⚠ DEPENDENCIAS HUÉRFANAS` — a dependency naming an id that exists nowhere.

Stop on any parser warning or orphan dependency and fix the spec (see
`orch-spec`) — either one usually means a task or a field silently went
missing. The one warning that is safe to leave is a `project_id` that does not
match the active project, and only when the user confirms the spec belongs
here; the id to use is the one on the first line of `orch explain`.

Things to point out to the user before they confirm:

- an **updated** task that is already `in-progress` or `done`: its runtime
  status and comments survive the merge, but its brief changes under it;
- orphans that look like a *renumbered* task rather than a removed one —
  renumbering loses history; restore the old id in the spec instead;
- a large number of new tasks in one go, which usually means the spec needs
  splitting.

## 2. Apply — only after the user says yes

```bash
orch atomize --apply                    # same scope as the preview
```

It writes `tasks.json` and a `tasks.json.bak-<timestamp>` next to it. Leave the
backups where they are; they are the undo.

Tasks created by atomize start as **`backlog`**. Existing tasks keep their
status, files and comments.

## 3. Make the DAG valid

```bash
orch router validate        # every task model resolves to a route?
orch validate               # schema, dependencies, cycles
```

**Both** have to exit 0. `orch validate` alone is not enough: when
`model_router.yaml` has no routes at all it skips the route check and reports
zero errors, while `orch router validate` still lists every unrouted task.

If a model does not resolve and the router already has routes, prefer fixing
the spec: change `Model` to one of its keys and run steps 1–2 again. When the
router is empty — normal for a project scaffolded without `--template` — or the
user really wants a new model:

```bash
orch router add-missing
```

It prints the entry it infers for each unrouted model (backend, `cli_model`,
tier) and asks `[y/N]`; anything but `y` aborts without writing. Show the user
those entries before they answer — a wrong `cli_model` only fails at dispatch
time — and remind them that every added entry gets tier `standard` unless
`--tier` says otherwise. `--yes` skips the prompt; do not pass it on the
user's behalf.

## 4. Promote what should run

orch only dispatches tasks whose status is **`todo`**, so freshly atomized
work does nothing until someone promotes it. Ask the user which tasks — often
the first phase, rarely everything — then, one per task:

```bash
orch task set --id F1.1.T1 --status todo
```

## 5. Report

Summarise: tasks added, updated and orphaned, what is now `todo`, and what
`orch explain` says is ready. Suggest the next command without running it:

```bash
orch explain                # what is ready and what is safe to do next
orch run --max-tasks 1      # a first dispatch, when the user wants one
```

Dispatching spends real tokens and opens PRs. It is the user's call, never this
skill's.

Attribution

hectorcanaimerohectorcanaimero
View sourceMore from hectorcanaimero →
SSkills DirectorySkills Directory

Ship a skill? Prove it's safe.

Free 120-pattern security scan, letter grade, and an embeddable README badge.

Submit a skill

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments (0)

No comments yet. Be the first to comment!

SSkills DirectorySkills Directory

Ship a skill? Prove it's safe.

Free 120-pattern security scan, letter grade, and an embeddable README badge.

Submit a skill

Related Skills

Caveman

Ultra-compressed communication mode that cuts output tokens while keeping technical accuracy. Levels: lite, full, ultra and the wenyan variants. Use for /caveman, "caveman mode", "talk like caveman", "be brief" or "less tokens".

1074701 votes

Hyperplan

Adversarial multi-agent planning skill. Self-orchestrates 5 hostile category members (unspecified-low, unspecified-high, deep, ultrabrain, artistry) via team-mode for ruthless cross-critique debate, distills only the defensible insights, then MANDATORILY hands the distilled insight bundle to the `plan` agent for executable plan formalization. Use when planning needs maximum rigor and surfacing of weak assumptions, blind spots, and over-engineering. Triggers: 'hyperplan', 'hpp', '/hyperplan', ...

693621 votes

Mcp Code Execution

Routes multi-tool workflows through MCP servers for large datasets and pipelines. Use when Bash tool overhead is limiting throughput on data-heavy tasks.

3351 votes

catchup

Recovers the conversation and failed tool calls of a previous Codex, Claude Code, Antigravity, Cline, Copilot CLI, Cursor, DeepSeek Harness, Kimi, OpenCode, Pi Agent, or ZCode session. Use when the user says "catch up", "what did the last session do", "get me up to speed", "I switched agents", asks to recover/summarize a previous session before continuing, or asks to diagnose or report a catchup failure. Do NOT use for the current conversation, git history, or any non-agent log.

691 votes

math-skill

A comprehensive mathematical reasoning skill for AI assistants — handles arithmetic to research-level problems with rigorous step-by-step reasoning, systematic verification, and transparent uncertainty handling

381 votes
View all in ai-agents →