Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add sergeyklay/.agents --skill delegate-workgit clone --depth 1 https://github.com/sergeyklay/.agentsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/sergeyklay/.agents/delegate-work)<a href="https://agentmods.dev/skills/sergeyklay/.agents/delegate-work"><img src="https://agentmods.dev/badge/skills/sergeyklay/.agents/delegate-work/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/sergeyklay/.agents/delegate-work"><img src="https://agentmods.dev/badge/skills/sergeyklay/.agents/delegate-work.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00143 | $0.03394 |
| Opus 5 | $0.00072 | $0.01697 |
| Sonnet 5 | $0.00029 | $0.00679 |
| Haiku 4.5 | $0.00014 | $0.00339 |
Grade A, and why
delegate-work scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 153 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Delegate Work and Verify What Comes Back
Your only view of delegated work is the delegate's own account of it. That account is written by the one party with an interest in it reading as complete, from a context you cannot inspect, and it is silent in exactly the places where it is wrong. A report that says nothing about instruction four is indistinguishable from a report whose author never received instruction four.
The defense is not skepticism, it is structure: make each instruction produce a named artifact, then check the artifacts instead of reading the prose.
research-it owns what a delegated conclusion is worth (a claim, not evidence). prove-checks owns whether a check could have gone red. This skill owns the orchestrator's side of the exchange: the brief, the delivery, and the report as a whole.
Trigger
- A prompt for a subagent, a background agent, or a parallel session is about to be written.
- Two or more agents will run against the same repository or worktree.
- A correction, a scope change, or a new instruction was sent to an agent that is already running.
- A returned report is about to be relayed to a user, quoted, or used as the input to the next delegation.
- The delegated work is expensive enough that re-running it is not the cheap fix.
Not this skill: measuring what a run cost (audit-agent), proving a green check was capable of failing (prove-checks), or deciding whether a technical claim the delegate made is actually true (research-it).
Procedure
1. Hoist the standing guardrail block out of the brief
Anything repeated into a second brief is written down once and referenced, never retyped. Four hand-copied blocks are four blocks that will drift, and the drift is invisible: each brief looks right on its own.
Keep one file (a shared brief fragment, a section of AGENTS.md, or a skill the brief names) covering the categories below. The instances are project-specific; the categories are not.
| Category | What it must settle |
|---|---|
| Isolation | Which worktree or directory the agent owns, and that it owns it alone. |
| Forbidden commands | The exact commands that would destroy a parallel session's uncommitted work. Name them; "be careful with git" is not an instruction. |
| Gate selection | How this project's checks choose the files they check, when that differs from "everything". |
| Evidence standard | Real command output pasted, not paraphrased. A summary of a gate run is not a gate run. |
| Negative control | Required before any green counts. Defer to prove-checks rather than restating it. |
| Incremental findings | Findings written to a file as they accumulate. Defer to the working agreement's reporting rule rather than restating it. |
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- yesterday Changed · +3 lines 7adb006bdca5
- 3d ago First seen · 150 lines · 143 tokens per session scan A a32947367a2d
delegate-work is a skill published in the GitHub repository sergeyklay/.agents (5 stars, last pushed yesterday), licensed Apache-2.0. It adds 143 tokens to every session and 3,394 once invoked, about $0.0007 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-08.
Other skills, from other repositories
chatdoc-studio-api
ChatDOC Studio API usage guide - complete documentation and examples for PDF parsing, chat applications, agent applications, content retrieval, and data extraction APIs.
war-room
Multi-agent research war room. Personas debate in sequential turns through two phases — ideation and proposal writing. Persona persistence and drift detection are enforced every turn via the persistent-persona skill.
system-info
Probe system resources — CPU, RAM, disk, GPU/CUDA/MPS — and translate findings into experiment design recommendations. Run this at the start of any compute-intensive project.
persistent-persona
Mechanism for keeping an agent anchored to a caller-defined persona across sessions. Covers the biplate private memory system for drift detection and self-correction.
memory-checkpoint
Defines how to save and load agent state as checkpoint files.
ChatDOC Studio--KnowledgeMate
Create and operate ChatDOC Studio knowledge bases through pdrouter using a Bearer API key and JavaScript helpers. Use when Codex needs to upload one or more PDF/DOC/DOCX files, skip failed files without aborting the whole job, create a knowledge base from successful uploads, or call the ChatDOC Studio knowledge-base…