delegate
/delegate
Use when work belongs somewhere other than this session (a different repository, another environment, or a visual design tool), or before creating or changing a skill, to decide where it lives.
- Author
- LandonSchropp ↗
- Repository
- agent-toolkit ↗
- Licence
- MIT
- Last repository change
- 27.10.2026
Assessment and security
- Files in this package
- 6
- Scripts
- 1
- HTML comments
- 0
- URLs in SKILL.md
- 0
No signals detected. Repository changed on 27.10.2026, without references to retired models or APIs.
MIT, open: use and modify while retaining author attribution.
- 1 scripts in this package. Review them before installing, as Claude may execute them.
Scripts: scripts/prompt.sh
What it does and when to use it
Use when work belongs somewhere other than this session (a different repository, another environment, or a visual design tool), or before creating or changing a skill, to decide where it lives.
Original author description
Use when work belongs somewhere other than this session (a different repository, another environment, or a visual design tool), or before creating or changing a skill, to decide where it lives.
How it works
- Choosing the Environment
- Skills
- Writing the Prompt
- Telling the User
- Rationalizations
SKILL.md preview
--- name: delegate description: Use when work belongs somewhere other than this session (a different repository, another environment, or a visual design tool), or before creating or changing a skill, to decide where it lives. --- # Delegate Hand a task to the environment best placed to do it. Choose the environment first; it decides everything else. ## Choosing the Environment Answer these in order and stop at the first that decides it: 1. **Is it engineering?** If not, it goes to [Non-Engineering](references/non-engineering.md). Visual design is Non-Engineering work done in [Claude Design](references/claude-design.md). 2. **Is it personal engineering?** It goes to [Local Engineering](references/local-engineering.md), since Cloud Engineering only takes work. `is-work-computer` tells you which machine you're on. 3. **Can success be verified without the user's judgment?** If so, and it doesn't need the user's Mac, it goes to [Cloud Engineering](references/cloud-engineering.md). If it needs the Mac, the development environment, or the user's judgment as it goes (naming, UX, architecture designed in real time), it goes to [Local Engineering](references/local-engineering.md). Route by the task, never by where this session happens to be running. Your own environment changes how you reach a target, never which target gets the task. When the questions don't settle it, ask the user. ## Skills A skill that lives in a repository goes to Local Engineering: the one it is already in, or else `personal-agent-toolkit` when it's about managing the user's personal life or shouldn't be in a public repository, `agent-toolkit` when it's the user's own engineering tooling, and `babylist/claude-plugins` when it's for work. Non-Engineering's own skills are built and maintained there rather than delegated. ## Writing the Prompt The receiving agent has none of this conversation, and nothing it replies reaches you. A prompt saying "turn what we just did into a skill" is worthless to it. State, in the task's own terms: - What the task is, in full. Where it came from a Linear issue, give the issue's URL rather than repeating its contents. - What already happened that the agent needs to know, and where: repository, files, commits. - What you have and have not already changed yourself, so it doesn't redo or undo work. Pass the task through as the user gave it. Don't reinterpret it, improve it, or decide it needs less than it asks for. Each environment's reference adds what its own prompts need on top of this. ## Telling the User Delegation is final. Say what was sent and where, then stop — never tell the user the agent will "report back" or "let you know" when it's done. Nothing it does flows back through this session; the user checks its progress by watching its own workspace or session directly. ## Rationalizations | Thought | Reality | | ---------------------------------------------- | --------------------------------------------------------------------------------- | | "It's a two-line change, I'll just do it here" | Then it lands without the target's conventions. Delegate it. | | "It's research, so it needs a workspace" | It edits no repository, so it is Non-Engineering work, not a checkout. | | "I'll wait for the agent's turn to finish" | The turn is the agent's, and its reply is the user's. Send it and move on. | | "I'll run the `design` skill instead" | It is the smaller preview. Claude Design work goes to claude.ai/design. | | "I'll reference what we just did" | The receiving agent was not here. Spell it out. | | "I'll reuse the running agent, it's faster" | A fresh worktree keeps the work on its own branch. Reuse only when asked. | | "I'll paste in the Linear issue's description" | Give the URL. The agent reads the issue itself; the issue is the source of truth. | | "The skill is only used by Non-Engineering" | It still lives in a repository, so it goes to Local Engineering. | | "I'll say it'll report back when it's done" | Delegation is final. Nothing reaches this session; the user checks the workspace. |
Files in this package (6)
- SKILL.md 4 KB
- references/claude-design.md 1 KB
- references/cloud-engineering.md 1 KB
- references/local-engineering.md 2 KB
- references/non-engineering.md 1 KB
- scripts/prompt.sh 4 KB
Versions
| Repository | Score | Licence |
|---|---|---|
| This page · LandonSchropp/agent-toolkit | 90 | MIT |
| aayushpokhrel1/delegation-pipeline | 100 | MIT |
| Filipeno/ai-hub | 100 | MIT |
| shakestzd/htmlgraph | 100 | MIT |
| shakestzd/htmlgraph | 100 | MIT |
| easeit-cz/vornik | 100 | AGPL-3.0 |
| hugrverk/riggverk | 100 | MIT |
| easeit-cz/vornik | 100 | AGPL-3.0 |
| aayushpokhrel1/delegation-pipeline | 100 | MIT |
Database snapshot: 13.11.2026 · package version 962ae6fcd11e (SHA-256 digest of the entire package)
Downloads in the last 30 days: 0 · ranking
Report a problem with this skill: harmful content, incorrect descriptions or an author’s removal request.
Similar skills
- 95delegate-claude
Delegate a bounded implementation, prototype, benchmark, or file artifact to Claude Code, then verify and integrate its completed work. Use when the user explicitly authorizes Claude to own execution. Read-only advice belongs to consult-claude. - 80cmk:delivery-handoff
Use when the user asks for "a handoff prompt", "a prompt for codex/grok/claude", "something I can paste into another agent", or wants to continue tracked work in a different tool — and at any phase boundary of the cmk:delivery-pipeline skill when the operator prefers a different… - 100background-agent
Use when a task is too long to block the current session or should run autonomously — delegates to a cloud background agent via & or /delegate so work can continue on GitHub while you keep moving locally, not in a local async PowerShell session. - 100delegate-to-subagents
Decide whether to delegate work to a subagent, and pick the right agent type and model when you do — so verbose output stays out of the main context instead of being paid for on every later turn. Use this whenever a task will produce output you will not reference again (test… - 100peers
If files change under you that you did not change, another agent may be editing the same repository; load this skill to find out before you touch them. Also load it before switching branches, rewriting history or stopping processes in a repository others may share, when the user… - 100delegate-work
Brief a background agent, confirm the brief arrived, and verify its report before acting on it. Use when writing the prompt for a subagent or parallel session, when several agents will run against one repository, when an instruction sent to a running agent goes unmentioned in…